X Prize Foundation Wants AI Physician On Every Smartphone

kkleiner writes "One of the exciting ideas being tossed around recently at the X Prize Foundation is the creation of an Artificial Intelligence physician that you could access from your smartphone. Want to know if that rash on your leg is poison ivy or smallpox? Take a photo of it with your phone and ask the AI. The possibilities are enormous, especially for the billion plus people around the world who live more than a few hours' walk or drive from the nearest doctor." This is one of four X Prizes in planning for the future. The other three are for an AI automobile driver, organ generation through stem cell use, and a deep sea submersible capable of exploring the sea floor.

Comments

No smallpox.

[Anonymous Coward]

Want to know if that rash on your leg is poison ivy or smallpox?

Not smallpox. C4n I plz haz the prize?

Re:No smallpox.

[Anonymous Coward]

Yo mama's so fat, she had bigpox.

Re:No smallpox.

[Impy the Impiuos Imp]

"Oh, I swore I would never use one of these. Activate cellphone emergency medical technicion!"

"Please state the nature of the medical emergency!"

"I'm a nerd and don't know what to say to this girl."

"I beg your pardon? I'm a doctor, not a song-and-dance Don Juan."

"I don't care. Tell her she's fugly but you might do her. Wave your arms and barf on her shoes. Anything. Just buy me enough time to run away!"

Internet hypochondria is already a phenomenon

[Burz]

http://www.webmd.com/balance/features/internet-makes-hypochondria-worse

Would the smartphone version be any better?

Re:Internet hypochondria is already a phenomenon

[hedwards]

Depends how it's done. If they manage to make it more of a tricorder eventually including some testing equipment or the photos are sent to somebody with knowledge hen probably not. It's not likely to replace an exam any time soon, but it would be helpful for contacting the consulting nurse as to whether or not to come in for an appointment.

But yes, if it just looks things up without any kind of smarts, then it's most likely just going to make things worse.

Re:Internet hypochondria is already a phenomenon

[Pojut]

Does anyone else find it hilarious that one of the primary contributors to Internet Hypochondria has an article about how bad Internet Hypochondria has gotten?

Re:Internet hypochondria is already a phenomenon

[commodore64_love]

>>>Would the smartphone version be any better?

"Please state the nature of the emergency. Oh. It's you again. What now???" ----- I suspect not any better. And then after people got sick of the AI Doctor, he would be relegated to mining duty.

Re:Internet hypochondria is already a phenomenon

[Monkeedude1212]

Well - wouldn't the smart phone version be able to differentiate the diseases from common blunders?

I could see a hypochondriac using this feature to learn that "That Bump" is just a pimple.

Which, I know - doesn't usually help hypochondriacs, even a real doctor generally doesn't sway their mind. So I don't see why a smartphone app when the internet is already widely available making it any worse. (If you can find more people who own a smart phone who don't have access to the internet at least once a week than I have fingers, I'll retract my statement)

Re:Internet hypochondria is already a phenomenon

[ColdWetDog]

Does anyone else find it hilarious that one of the primary contributors to Internet Hypochondria has an article about how bad Internet Hypochondria has gotten?

No, it just gives me the heebie-jeebies. And a headache.

Re:Internet hypochondria is already a phenomenon

[IICV]

Exactly - the problem isn't creating an AI physician, the problem is creating an AI physician that you can't lie to blatantly. It would be trivial to create a little program that's little more than an enhanced version of those 20 question balls and runs a differential diagnosis engine; something like that would cover 90% of all diseases after a few rounds. It could show you little pictures like "do your bumps look more like this or more like this?", or any number enhancements. It would probably work incredibly well with honest, impartial patients.

The problem is, it has no way of knowing if you're lying. If you say "yes, when I press hard on this point on my belly it hurts when I let go", the thing has no way of knowing if you pressed on the right place or if it actually hurt. If there's a doctor doing that to you, they'll have a pretty good idea if it hurt.

Really, the best this thing can do is either say "you should see a doctor" or "you should take two acetaminophen and ask me again tomorrow"

Re:Internet hypochondria is already a phenomenon

[Icarus1919]

Actually, this would be perfect for all that hypochondriac that lives more than a few hours walk from a doctor in a third world country, but has a smartphone.

Re:Internet hypochondria is already a phenomenon

[kg8484]

No, but now I'm worried I might have Hypochondria.

Re:Internet hypochondria is already a phenomenon

[jedidiah]

This is your usual example of an expert system from an undergraduate introductory AI course.

I'm not sure such a system warrants an X prize. This is the sort of thing that isn't done more because of legal liability than technical limitations.

Although a photographic "name that rash" app could be cool.

There are already websites that you can use for fueling your hypochondria based on a list of symptoms.

Re:Internet hypochondria is already a phenomenon

[Anonymous Coward]

It's worse. Imagine combining it with these technical challenges.

If it's a significant technical challenge to even figure out what part of the body is in a photo, imagine the challenge of trying to diagnose a medical problem with such a photo.

Re:Internet hypochondria is already a phenomenon

[elrous0]

Don't worry, the AMA would never allow it (not in the U.S. anyway). Their main function is to protect the livelihood of their members (aka physicians). Anything that threatens their monopoly is immediately labeled a health hazard and banned in the U.S.

Re:Internet hypochondria is already a phenomenon

[imakemusic]

Now that you mention it, I think I've got a headache, too...

Re:Internet hypochondria is already a phenomenon

[fuzzyfuzzyfungus]

Based on the size and influence of the America "nutritional supplements" industry, the ADA won't be able to do much more than force US sellers to print "This product is for novelty purposes only and is not intended to treat, diagnose, or cure any disease. This statement has not been evaluated by the FDA." somewhere on the package.

You can sell virtually anything, so long as you slap that on the package and stick to vague claims like "improves immune function" and "enhances wellbeing". You don't even have to comply, in any serious way, with purity and potency requirements. Only if it is reported to the FDA(they don't proactively investigate) that your product is injuring or killing people, will they come after you. The FTC might come after you if you make too many blatantly false statements in your commercials; but that is a distinct problem.

Re:Internet hypochondria is already a phenomenon

[D.A. Zollinger]

In the industry, this is a concept that has already been talked about quite a bit. In fact, a common mantra heard at AMIA conventions is the oft recited, "Any physician who can be replaced by a computer should be."

Unfortunately, that is MUCH easier said than done. And while clinical decision support systems exist to help physicians with their patient diagnosis, every physician uses them as a guideline, and not as an absolute reference or comprehensive source of information. Experience is heavily valued, which is why every physician has 7+ years of postgraduate education and working in the field.

Re:Internet hypochondria is already a phenomenon

[delinear]

I hope it's not contagious, I'm bound to catch it, I seem to catch everything, I'm convinced I'm always ill. What are the symptoms?

Re:Internet hypochondria is already a phenomenon

[Pence128]

It's brain cancer.

Re:Internet hypochondria is already a phenomenon

[Burz]

I thought it was pretty funny too. But it's not ironic or hypocritical... they're recognizing how their site can sometimes have a negative influence.

Re:Internet hypochondria is already a phenomenon

[Kilrah_il]

It's even more problematic than you think. As a doctor, many times I have patients coming to me with "agonizing abdominal pain" - they are sure it's appendicitis. If you check their stomach or ask them "does it hurt here?" they jump and cry and wail, etc. etc. But if I start talking to them on other subjects (What do you do in your life? Are you married? Children?) and get them diverted, I find out many times that they "forget" about their pain and the stomach is as soft and non-tender as can be.
This is one of the reasons that no app can replace a physical exam by a doctor. You need the doctor-patient relationship to strip away the anxiety and find out the true magnitude of the symptoms. So, yes, an app like the one in TFA could be nice as a handy reference, but nothing beats the good-old face-to-face meeting.

Re:Internet hypochondria is already a phenomenon

[Anonymous Coward]

The AMA doesn't give a damn about it's members. Like most organizations that size, it only cares about perpetuating itself.

Re:Internet hypochondria is already a phenomenon

[bitflip]

I was having some pain in my chest a couple of years back, and did some research on the internet, but none of the symptoms fit. So, I went to my doctor, and told him I'd done some research, but it didn't seem like a heart problem. When he looked surprised, I asked him if I was the only person who'd said they'd looked stuff up on the internet, and decided nothing was wrong. He said, pretty much, yea.

(turned out to be a minor stomach problem, all fixed now - thanks doc!)

Re:Internet hypochondria is already a phenomenon

[imakemusic]

According to my phone it's appendicitis.

Re:Internet hypochondria is already a phenomenon

[Anonymous Coward]

It's not a tum-ah!

Re:Internet hypochondria is already a phenomenon

[Anonymous Coward]

Can't this be said about every job and every person who plays the U.S. cotton paper game?

Re:Internet hypochondria is already a phenomenon

[DriedClexler]

Oh, yes, the way that AMA members shut down the self-diagnosis interface on WebMD was just *horrendous*. Oh, wait...

Look, I hate the AMA, and I hate the way doctors impede meaningful improvements in the process of healthcare. But seriously, they don't worry me here. Even if they don't like this program when it gets invented, it will probably not be banned, as long as it has some disclaimer like:

"Despite the 99.9999999% success rate of this device, you should completely distrust these results and go see a real doctor because this shit hurts our profits. Carry on then!"

Re:Internet hypochondria is already a phenomenon

[jeffporcaro]

OMG, I *wish* this were true of the AMA. As a physician and still active member, I can tell you that this couldn't be further from the truth. The AMA's primary business is publishing and maintaining insurance coding and billing standards, and selling their databases to the highest bidder. They employ lobbyists primarily to maintain that monopoly - they are NOT particularly interested in maintaining insurance or government payments to physicians (aka "livelihood"), although they make noises on that topic occasionally. They've basically been relegated to the sidelines on most national issues involving medicine. They represent less than 30% of active US physicians. I hear this same trope frequently, however, despite the fact that it's demonstrably false.

Re:Internet hypochondria is already a phenomenon

[Scrameustache]

make it more of a tricorder eventually including some testing equipment

I see a great potential for cheap, on-the-shelf testing kits made to create visuals clear to the machine. They could become as ubiquitous as band-aids.

Re:Internet hypochondria is already a phenomenon

[thesandtiger]

It doesn't even have to be dishonesty - it can be a misunderstanding.

If you say "does it hurt when you press on this spot" there are a LOT of assumptions: what does "hurt" mean, really? How much pressure in that spot? Point pressure or spread over a slightly larger area? Hurt *where* - inside or out? What kind of hurt - sharp, stabbing pains or a dull ache or something else? How intense is the pain? What if you're a giant sissie and any kind of pain makes you shriek? What if you're made of cast iron and even something that would cause 99% of the population to pass out is just something you shrug off?

Re:Internet hypochondria is already a phenomenon

[NevarMore]

As a patient, how do I find a doctor that I can actually have a good working relationship with?

I had a good one growing up, saw him through high school and college. Its not practical to travel 2 hours to see him now. How do I "feel out" a doctor for the first time? Even scheduling a physical involves a time and money investment on my part.

As a doctor I assume this benefits you, a generally healthy patient who follows directions, values your opinion, and pays his bills has got to be better than a prescription shopper or a nutter.

Re:Internet hypochondria is already a phenomenon

[kklein]

That might be the difference between nerds and normals. When I do medical research, mostly I figure out what it's not. If it's still bugging me, I go in and say, "Here are the symptoms, but I read that it could be something or could be nothing, so I decided to get it checked out." When he says "it's nothing," I say, "Thanks! I feel better now."

I think part of the problem is that people often have a hard time fathoming how much more someone can know about something than they. I'm a college prof, so I think I've gotten used to feeling stupid--I know a lot more about my field than Joe Sixpack, but I can give you a 2-page list of people I know or know of who know more than me. And then there are the profs in other departments. Even the least-experienced ones know more about their fields than I ever, ever will.

I respect experts. I understand that mistakes happen, but they have a much better chance of knowing that I do.

I once had an anxiety attack, but I didn't know why. I was worried that there was something wrong with my heart, since there are heart problems in my family. After being checked out, my doctor very carefully and with a clear sense of dread, said, "I think the problem might have been... Mental." He watched my face for my reaction. I said, "So it's all in my head?" "I think so." "Thank god. I can control that."

We both felt relieved. Me, because my heart was fine. Him, because I didn't take offense at being told that I had just freaked out.

Re:Internet hypochondria is already a phenomenon

[Anonymous Coward]

I was having some pain in my chest

turned out to be a minor stomach problem

If your stomach is in your chest, I'm pretty sure that's more than a minor problem...

Re:Internet hypochondria is already a phenomenon

[CarpetShark]

I think I've got a headache, too...

I think you just have sexual problems.

Re:Internet hypochondria is already a phenomenon

[Kilrah_il]

I know it was a joke, but there is such a thing as referred pain - a gastric ulcer (AKA minor stomach problem) can cause chest pain; and a heart attack can cause upper abominal pain.

Re:Internet hypochondria is already a phenomenon

[Kilrah_il]

How do you find a good car mechanic? You ask around with friends, go to a few different mechanics and use your gut feelings and experience. No shortcuts. Sorry.

Re:Internet hypochondria is already a phenomenon

[kauttapiste]

No, it's Lupus.

Re:Internet hypochondria is already a phenomenon

[CarpetShark]

As a doctor, many times I have patients coming to me with "agonizing abdominal pain" - they are sure it's appendicitis. If you check their stomach or ask them "does it hurt here?" they jump and cry and wail, etc. etc. But if I start talking to them on other subjects (What do you do in your life? Are you married? Children?) and get them diverted, I find out many times that they "forget" about their pain and the stomach is as soft and non-tender as can be.

Be careful of this. On occasion, I have gone to a doctor and explained a subtle symptom, only to have nothing found, and then had to go back months later when the problem had been allowed to progress to the point where a doctor could clearly see it. That is NOT the way that doctors should work. If a patient complains, a proper investigation should occur until the problem is resolved.

Also, lately, I've seen a few doctors and related a host symptoms (I'm not a hypochondriac, just happen to (apparently) have a few related so-called immune disorders). The problem is that Doctors want to treat the major issues with some painkiller, and ignore the minor ones. Taken together, they are all symptoms of a bigger problem, which deserves a fuller investigation, but doctors want to treat the things that raise immediate alarm bells; the things they see every day and think they have quick fixes for. Except, those things are symptoms, not the issue, and they want to just keep presribing pills while a larger problem progresses uninvestigated.

Doctors really need to do more medical research. They DO NOT have the answers they think they have. Which is why so many are turning to the internet for answers.

Re:Internet hypochondria is already a phenomenon

[Kilrah_il]

I agree that sometimes a minor complaint is only the symptom to a major issue, and further, I agree that some doctors (I would call them Not-So-Good-Doctors) are quick to discount the patient's complaint and give him something to get rid of whatever pains him. OTOH, most of the people that have a small complaint don't have a big/major disease behind it. Do you want us to perform a CT for everyone with a stomach-ache? Apart from the cost, there is the issue of radiation exposure - I don't think you would want to be exposed to so much radiation for something that is almost always not serious. Perform a cathatarization to everyone with chest pain? I think more people would die from complications of the procedure than would be saved from having not missed a few heart attacks.
So the solution? Listen to the patient, perform a meticulous examination and follow-up on the patient. One thing that recurs many times at presentations of rare/special cases is the need for repeat examinations. Symptoms that appear mild at the inital examination may evolve after some time has elapsed.
And yes, many doctors need more humility.

Re:Internet hypochondria is already a phenomenon

[Ravenscroftj]

Agreed, I recently had an onset of Internet Hypochondria which gave me actual physical symptoms, once I managed to calm myself down, I realised that the symptoms were all but gone!

Hmmm

[Spad]

Because having medical advice available on the internet hasn't led to people flooding GP surgeries because they're convinced their cough is actually Ebola.

Re:Hmmm

[hedwards]

That's why some health insurance companies provide a complimentary consulting nurse. Basically somebody that's available around the clock to answer those, should I call 911, go to the emergency room, make an appointment or just ignore it suggestions. Obviously they tend to be a bit action biased as doing a screening over the phone isn't easy, but it does help people make better decisions about what is urgent and what can wait.

Re:Hmmm

[akadruid]

In the UK we have NHS Direct which is both 24-7 telephone and online screening and advice, run by the national healthcare system. From personal experience, it's very good.

Re:Hmmm

[commodore64_love]

I hope they are better than the stories I've come out of NICE (aka nasty). Over in the State of France they have a screening service, where doctors answer emergency calls directly and can provide medical care immediately, or else send an ambulance if the sick person needs hospital treatment. It has saved the State Government a lot of money by eliminating un-necessary ambulance/hospital visits.

This seems a good idea for the US Member States to copy, hiring actual doctors to handle 911 calls, rather than some minimum wage person. If I was a politician I would shove it through the Legislature, but of course I'm a nobody say I have no say whatsoever. Oh well.

Re:Hmmm

[vlm]

consulting nurse

Somehow, despite it "requiring" AI, for liability reasons, other than the most trivial of follow up issues, it'll probably end up 99% of the time as a semi-realtime nifty videoconference frontend for existing consulting nurses.

I would also anticipate some "self nursing". Rather than paying someone to glance at a sutured wound, have the patient photograph it with their cellphone and have a centralized nurse (and/or dr) review a stack of pictures at once. Rather than paying a nurse to stick a thermometer in the patients mouth (or, where ever) and wait for a result and record it, the phone might bug the patient to test himself and log his own result.

This idea of a personal universal electronic medical record is kind of interesting, compared to the corporate electronic medical records we are "sort of" seeing deployed.

Re:Hmmm

[Scrameustache]

Because having medical advice available on the internet hasn't led to people flooding GP surgeries because they're convinced their cough is actually Ebola.

Right, the internet created hypochondriacs, just like videogames created murderers and the Austrian waltz created coveting of ankles.

Re:Hmmm

[ultranova]

This seems a good idea for the US Member States to copy, hiring actual doctors to handle 911 calls, rather than some minimum wage person. If I was a politician I would shove it through the Legislature, but of course I'm a nobody say I have no say whatsoever. Oh well.

The problem is, it would cost tax money and be providing a public service, and collecting taxes to provide a public service is against the oh-so-fashionable right-wing ideals.

Re:Hmmm

[Yvanhoe]

Medical expert on internet is usually of the kind "Hey! it looks like what I had! You should try [random cure]". The proposal is for a system expert that would have an exhaustive medical knowledge of the possible causes and proposing probable cures. It could actually beat some human doctors. Remember the game of 20 questions ? Computers are stupidly good at that. Recognizing symptoms of a disease is the same kind of game.

"-I have a red harsh on my leg"
"-Do you have some on both legs ? Y/N"
"-Did you get in contact to grass recently ? Y/N"
"-Do you have a known allergy to sleptigiornocius ? Y/N"
"-Do you have a known allergy to synthetic clothes ? Y/N"
"-Are you currently on Cyclomosil treatment ? Y/N"
Most probable cause : Cyclomosil allergy.
Second probable cause : Eczema.

Why couldn't it work ?

Re:Hmmm

[delinear]

[...] it'll probably end up 99% of the time as a semi-realtime nifty videoconference frontend for existing consulting nurses.

Sounds like Chat Roulette for the medical sector. Hmm, insert your own jokes here.

Re:Hmmm

[BeardedChimp]

I'll vouch for NHS direct, my Dad wrote the flow charts for it.

The original was written using templates in word97 I'm guessing that was pretty painful.

Re:Hmmm

[spire3661]

Using a doctor to handle 911 calls is an incredible waste of resources. 911 handles alot more situations then just medical. Let doctors tend patients and let low skill workers man the phones. Hell, if you want to have a single doctor in the call center, thats not a terrible idea, but still a waste of resource.

Re:Hmmm

[camperdave]

You can't take the sky from me...

Dangit! Now I have the theme music stuck in my head.

Re:Hmmm

[Scrameustache]

Rather than paying a nurse to stick a thermometer in the patients mouth (or, where ever) and wait for a result and record it, the phone might bug the patient to test himself and log his own result.

I'm sorry to say that you're grossly overestimating the reliability of the general public. Ask something simple like temperature measurements and people will rinse their thermometer in hot water to clean it, or ice water, or spray rubbing alcohol on it after taking it out but before logging the results, etc. Ant that's not even the creative ones; you'll get stories involving goats and pressure cookers within weeks of launch.

Re:Hmmm

[commodore64_love]

No actually it isn't. Right-wing (meaning small government supporter) doesn't object to STATES exercising various powers, such as regulating the roads and providing 911 services.

Plus having doctors on the line who could treat medical conditions immediately, instead of sending an ambulance to every person who has a minor ailment (a cold or scraped knee), would SAVE money over the long term. No need to raise taxes.

Re:Hmmm

[commodore64_love]

>>>Using a doctor to handle 911 calls is an incredible waste of resources.

I thought I was speaking in English. Was I not speaking in English? *In my last post* I clearly said that France does exactly that (doctors handle 911 calls), and it's saved them a shitload of money. Of course these French doctors only handle the medical 911 calls, not the fire or police calls, but still it sounds like an effective, cost-saving system.

Re:Hmmm

[HiThere]

Nothing wrong with sticking it in rubbing alcohol after using it. At least if it's one of the old mercury ones. They had a marker that floated on top of the mercury column and would stay in place when the thermometer cooled off. That's why you had to shake it down.

(I'm having a lot of trouble, though, imagining the story about goats that you heard.)

Al Physician

[lowrydr310]

I read the title as Al (like short for Albert or Alfred) Physician - I thought to myself, "What a funny name for a doctor."

Dang sans serif fonts...

Re:Al Physician

[Scrameustache]

Al Physician is the healthplan of Al Queda.
"Just blow yourself up with this fertilizer and diesel and call me in the morning if the rash is still there."

If it's smallpox, there's a bigger problem

[Anonymous Coward]

Since smallpox is the only disease the human race has ever eradicated, an outbreak would be a very big deal. Unless you work in a germ warfare/infectious disease lab, in which caseyou'd already either know what it looks like or have access to people that do.

My epitaph...

[elrous0]

Here Lies Jim
His cellphone said it wasn't cancer.

Re:My epitaph...

[javilon]

Here lies Joe, he had no medical insurance and the only available means for him to get some (admittedly bad quality) medical diagnostics on his melanoma where forbidden by the bureaucracy for "safety" reasons.

Re:My epitaph...

[Sentrion]

And that is why our health care is so expensive. If sub-standard providers were allowed to practice there would be more mal-practice but there would be more options for those who don't have any care at all. Medicaid and charity care are too spotty to cover everybody who needs affordable care, not to mention the many exclusions. Our current regulations only exist to protect the upper middle class who are too stingy to pay top dollar for quality care. Afterall, they have yacht payments to worry about, and who has time for background checks? For now the best your smart phone will be allowed is a Voodoo app.

Re:My epitaph...

[Sentrion]

Don't worry, the new Smart Phone, M.D. will come with a label: "Warning! This phone emits radiation known to the State of California to cause cancer ... but this phone will text you to let you know when it happens."

Bad idea

[Kazymyr]

A smartphone cannot perform a physical exam. Enough said.

Re:Bad idea

[Anonymous Coward]

I've seen evidence that they can do a gyno exam though...

Re:Bad idea

[pseudorand]

And if it could, I think we could write much more popular apps than an AI Physician. ;)

Re:Bad idea

[sznupi]

Not available medical personel can't perform a physical exam, either (but a phone can connect one with existing personel, if that's required)

Re:Bad idea

[Anonymous Coward]

And prostate exams would be a bitch.

Re:Bad idea

[bill_mcgonigle]

A smartphone cannot perform a physical exam. Enough said.

You should write the X-Prize folks about this. They may not realize this and wind up releasing an AI app that can't properly determine that a physical exam is required for additional data.

Re:Bad idea

[M8e]

I wonder if an AI physician could make an oral exam.

Re:Bad idea

[Kazymyr]

Moot point. A physical exam is required for pretty much [i]any[/i] diagnosis.

Re:Bad idea

[Kazymyr]

Oops [] <> :)

Re:Bad idea

[buchner.johannes]

I know! They should build a phone that allows you to talk to people at a different place!

Re:Bad idea

[CarpetShark]

A smartphone cannot perform a physical exam.

No, but the vibrate mode might be good.

The possibilities are enormous...

[Angst Badger]

...for an epidemic of medical student's disease.

Tell me more

[bsDaemon]

Tell me more about X Prize foundation Wants AI Physician On Every Smartphone. /emacs

Re:Tell me more

[Anonymous Coward]

Yow! Legally-imposed CULTURE-reduction is CABBAGE-BRAINED!

Nice idea, but...

[miaDWZ]

The thing is going to get killed a week after public release after the AI 'misdiagnosis' someone and they decide to sue.

Re:Nice idea, but...

[Anonymous Coward]

Yeah because doctors never misdiagnose anyone right?

Re:Nice idea, but...

[Max Romantschuk]

There is a world outside of the litigation-happy US, you know...

Re:Nice idea, but...

[Anonymous Coward]

In the US they can slap the label "diagnosis for entertainment only" on the phone.

Re:Nice idea, but...

[blackraven14250]

More likely to be on the splash screen for the app, but the idea is there...

Re:Nice idea, but...

[elrous0]

Yeah because doctors never get sued and don't have to carry malpractice insurance, right?

Re:Nice idea, but...

[fuzzyfuzzyfungus]

The difference is, "AI" systems are neither advanced enough, nor sufficiently embedded in the public mind to be seen as distinct from their creators.

If my doctor fucks up, I sue the doctor(and possibly the hospital, if they really should have known about his habit of bringing a hip-flask into the surgical suite). I don't sue the medical school that "produced" the doctor.

If an AI expert system fucks up, I don't sue the AI, I sue the company that built the AI. Thus, unless the company wishes to carry some sort of novel "aggregate malpractice insurance" covering all the products it sells, selling that product would be wildly uneconomic.

The same basic conditions apply with the various proposals to automate cars. Humans are shit drivers. It is easy to believe that machines will be able to do better within the decade, and do better cheaply not too long thereafter. However, courts have an easy time with the idea that humans suck, and are morally responsible. Therefore, unless a definite defect in the vehicle is detected, the driver is usually blamed. Even if AI drivers cut vehicular crashes by 80%, saving thousands of lives a year, the companies building them would be sued into oblivion.

Re:Nice idea, but...

[jdgeorge]

There may be such a world, but I don't live there.

If this is billed as more than a physician-assistance tool, it is opening itself up to horrific misuse by, say, "the common clay of the new West... you know.... morons."

Re:Nice idea, but...

[OutOfMyTree]

Just because the US has the world's most advanced system for suing random people, that will stop the development of a useful tool for people living far from medical help?

Why stem cells specifically?

[MozeeToby]

Why mention stem cells specifically in the prize description? I'd rather see something like "Create replacement organs in the lab with MTBF of X years" but I guess that doesn't have quite the same ring to it. Different organs lend themselves to different replacement strategies better than others. The first long term implantable artificial hearts are just coming to the market, at the same time stem cells are being used to build the first replacement bladders, also at the same time some basic nanotech is looking at replacing the pancreas. So... why call out stem cells specifically when the future is probably some fusion of many different approaches depending on need, cost, and ease?

Re:Why stem cells specifically?

[davaguco]

I completely agree with you, I have read about many ways to create and replace organs that are being studied, and a lot of them do not involve stem cells.

Re:Why stem cells specifically?

[damien_kane]

IANAPHD, but I'm guessing it's due to the studies showing that the body has a natural tendency to react against foreign objects. Rejection of the donor organ is one of the major concerns in transplants today (beyond the actual surgery), if I'm not mistaken.

A long-term implantable artificial heart is quite foreign.
A heart grown from one's own stem-cells, carrying one's own DNA within its cellular makeup is slightly less-foreign (even if it was grown in a lab instead of inside a womb).

Slacking

[Coffee Warlord]

Six posts in and no comments about your phone saying "please state the nature of your medical emergency".

Incredible.

The Doctor Says

[GreggBz]

Perfect. All you'll need is one of those phones with a built in projector.

"Please state the nature of the medical emergency."

Re:The Doctor Says

[Anonymous Coward]

I last saw Robert Picardo on Stargate Atlantis.

Obligatory...

[smooth wombat]

"Please state the nature of the medical emergency."

Great idea

[Black Parrot]

Lawsuits will be a problem. And though we've got the technology for AI medical diagnosis (at least for some stuff), the visual processing suggested by the story is still a bit beyond us. (Cf. yesterday's story about identifying images of genitals on chatroulette.)

Re:Great idea

[timeOday]

"A bit beyond us" is exactly what the XPrize is for.

Re:Great idea

[spire3661]

This is an idea that should be shielded from lawsuits. Tired of human innovation being held up because the people on the slow bus.

Pissing your money away

[Anonymous Coward]

I think this is one example of techno-utopian overkill.

I think you could spend much less money just getting more meat doctors into the bush.

I think this over estimates the ability of AI and anything limited to pure visual inspection could be accomplished by emailing a picture to a doctor using the phones we have now.

Now if this smart phone could do blood tests, you might have something.

Re:Pissing your money away

[colinrichardday]

I think you could spend much less money just getting more meat doctors into the bush.

But how gynecologists do you need?

fun thought

[tangelogee]

I would totally buy it if it said "Please state the nature of the medical emergency."

dr phone

[Anonymous Coward]

Maybe its just me, but if your have to walk miles to a Doctor, maybe having a smart phone isnt the best use of your money?

"AI"

[trifish]

Seriously, isn't it time to stop overusing, abusing and misusing the term AI? Such primitive software doesn't come even close to the kind of intelligence people with IQ of 100 or higher have. Thanks.

Re:"AI"

[bcmm]

Image recognition is certainly an considered a part of "AI" research, because it is still a task which is very simple for humans and very hard for machines. The term "AI" is a bit strange anyway - whenever a goal of AI research is reached, we realise that that task didn't' actually need "intelligence" (whatever that is) - for example, the world chess champion is a machine, and we still don't have consider AI to have arrived. I suspect we won't really notice the point at which machines start to get better at most jobs than humans.

Re:"AI"

[John Hasler]

> Image recognition is certainly considered a part of "AI" research...

According to the anti-AI crowd AI is whatever it is that computers can't do yet. There was a time when all agreed that a machine playing a credible game of chess would constitute proof that AI had arrived but now defeating grand masters is "mere computation".

Re:"AI"

[jedidiah]

...yeah. There was a time: the 50's perhaps.

Even before there was a computer big enough to defeat a chess grand master by brute force, such an undertaking would have been described as a weak example of AI.

Re:"AI"

[Anonymous Coward]

The reason for that is that when the problem was first approached, chess wasn't thought to be a game of computation, it was thought to be a game of guile, cunning and strategy. When the IBM researchers solved the problem with the Blue Gene machine, it was only then that the chess fraternity realized that the computer simply broke the game down into a sequence of decision trees,and played the most likely path at any given point. This really is "mere computation", and in no way constitutes intelligence by any sensible definition of the word.

It can play a game of chess, but it can't answer you when you ask it "What colour is my hair?"

No, AI has not arrived, and personally, I never think it will. At least, not in our lifetimes. The complexity of the human mind and the way that it interacts with the world is so monstrously complex that all I think we'll ever be able to do is simulate various aspects of it. We *may* one day simulate intelligence convincingly, but it'll never be convincing enough that we need the likes of Voight Kampff tests.

True AI belongs in the realm of anti-gravity machines and other sci-fi tech so wonderfully far away that the only real use they have is as plot-crutches in movies.

Re:"AI"

[Anonymous Coward]

Even if a program is able to recognize an image, the program still does not have intelligence. That's the point.

Re:"AI"

[Myopic]

There was a time when all agreed that a machine playing a credible game of chess would constitute proof that AI had arrived

When was that? In my college CS education we were taught that the Turing Test has always been the standard set for artificial intelligence. To me, being able to select the probably best chess move out of hundreds of millions of moves is, in fact, "mere" computation. (Really impressive computation, but not intelligence.) But I don't want to be painted as anti-AI.

Re:"AI"

[Lord Ender]

In modern Computer Science, AI refers to the class of software problems which come naturally to humans but not to computers. Adding large sets of numbers is not AI. Image recognition is AI. Image recognition is needed to diagnose illness via a webcam. "Expert systems" are also AI.

"IQ of 100" doesn't enter into it.

Re:"AI"

[Anonymous Coward]

That's nonsense. Artificial Intelligence means... uh, wait for it -- ARTIFICIAL *INTELLIGENCE*!

An algorithm that detects geometric shapes IS in fact primitive (when created), completely natural to computers, and does not create any kind of artificial intelligence.

Re:"AI"

[Lord Ender]

Spoken like someone who has never studied AI and hasn't even considered the definitions of the words he used.

Hmm, maybe I should write my own AI doctor app

[NotSoHeavyD3]

Of course I wouldn't want to get sued so my app would just tell you to go see a real doctor you idiot no matter what happened.

Re:Hmm, maybe I should write my own AI doctor app

[zmollusc]

Round here, you could replace the General Practitioner with a similar app that sent you to the hospital.

Yeah...

[SolitaryMan]

Especially for the billion plus people around the world who live more than a few hours walk or drive from the nearest doctor.

Yeah, to all four of them, who actually have a smartphone

Re:Yeah...

[delinear]

Maybe it's a prelude to the One Smart Phone Per Child initiative.

Re:Yeah...

[matt4077]

It'll take about two years for every phone to be a smartphone. And if you think the developing world doesn't have cell phones, you're wrong. Africa isn't little children with big hungry bellies anymore, at least not for the majority. Is simply skipping the landline and going directly for cell phone. Millions of people are using it daily, to a point where prepaid credit has become a substitute for banking. There's at least one cellphone in any group of twenty people, and that would be more than enough to always have one around for medical purposes.

Re:Yeah...

[mathimus1863]

When I visited Thailand, I was amazed to see that so much of the country is poor and without healthcare, yet they ALL have cellphones. In fact, Thailand has the 5th highest cellphone ownership rate in the world (1.25 cellphones per person, on average). It's crazy to go to a hill tribe village 2 hours from anywhere else, see that they probably don't even have running water, yet they're all listening to music or chatting on their cellphone. I don't even know how they charge them! I didn't think they had running electricity...

Re:Yeah...

[Anonymous Coward]

http://www.itnewsafrica.com/?p=2839

tl;dr "market penetration of cellphones in Africa is sitting at 37%, ... the continent has largely bypassed fixed-line telecommunications solutions in favour of mobile technology"

Now, I'm sure that a lot of that 37% resides in urban areas, but as service coverage continues to expand, I can see this being a valuable tool.

Re:Yeah...

[Anonymous Coward]

No, you have it all wrong...The billion plus people don't all have to get a smart phone. Only the doctor has to get the smart phone and then....oh wait...never mind...bad idea. Move along.

Re:Yeah...

[weirdo557]

no running electricity? then it had to be static /rimshot/

Re:Yeah...

[sznupi]

http://commons.wikimedia.org/wiki/File:Uganda_mobile_phone_charging_service.jpg

One of quite a few possibilities...

Re:Yeah...

[sznupi]

Opera State of the Mobile Web, November 2009, focusing on Africa. And just by glancing at lists of top mobile phones in listed places - there's at least one, sometimes few "pure" smartphone types on most. Where there aren't any, some of "feature phones" are pretty close anyway (many with multitasking BTW, something some of the "smartphones" can't do)

Similarly in Southeast Asia, or Latin America

Re:Yeah...

[sznupi]

...or only a small part of given community.

The diagnosis is...

[AdmiralXyz]

"You have severe inflammation of the cerebral cortex, human. The only cure is to wire your brain into the AI Overmind. Proceed at once to the nearest Community Conversion Center."

Privacy Concerns Abound

[Eric_ColonSlashSlash]

They need to pay particular attention to privacy concerns if the phone were to have any gathered physiological data or even self described symptoms - I wouldn’t want that information hacked or even disseminated under the best of intentions.... “Look – he’s having a heart attack or early onset diabetes – let’s cancel his healthcare [insurers] or blackmail him [hackers]”

Re:Privacy Concerns Abound

[joh]

They need to pay particular attention to privacy concerns if the phone were to have any gathered physiological data or even self described symptoms - I wouldn’t want that information hacked or even disseminated under the best of intentions....

“Look – he’s having a heart attack or early onset diabetes – let’s cancel his healthcare [insurers] or blackmail him [hackers]”

I think if Google would look at all health-related search terms they'd have a good base for that even now... and maybe they already do.

Re:Privacy Concerns Abound

[sznupi]

Seriously, you think it's plausbile that would stop people from letting a mobile / etc. monitor their health?
Oh, the wonders of corporate medical system?...

Re:Privacy Concerns Abound

[Dragonslicer]

I think if Google would look at all health-related search terms they'd have a good base for that even now... and maybe they already do.

There's no "maybe" about it. Google tracked flu-related searches this year.

A few hours walk from the nearest doctor.

[Anonymous Coward]

The nearest doctor is a several hours away walking. I don't have a car and there are no roads. But thanks to the AI software on my smartphone and ubiquitous 4G coverage, I have quality health care at my fingertips.

Re:A few hours walk from the nearest doctor.

[delinear]

On the downside, if the nearest doctor is two hours walk away, and the phone tells you you're going to die without immediate medical care, the stress might just kill you.

Re:A few hours walk from the nearest doctor.

[damien_kane]

If not, the lack of immediate medical care probably will.
I wouldn't worry too much about it, then.

to bad high data rates will kill this and roaming

[Joe The Dragon]

to bad high data rates will kill this and roaming fees can make seeing a real doctor cost less.

How many people...

[tompaulco]

How many people in the world that live several hours walk from a doctor or hospital have smartphones? How well does that smartphone work with no coverage? I don't think they tend to put in cell towers in areas where the nearest human population complex is 30 miles away.

Re:How many people...

[John Hasler]

> I don't think they tend to put in cell towers in areas where the nearest
> human population complex is 30 miles away.

ROFL. I live 30 miles (a 15 minute drive) from the nearest hospital. We have excellent cell phone coverage.

Re:How many people...

[3dr]

Patient: I've walked for six hours to get to your clinic, I'm famished and have severe abdominal pains. I need to recharge.

Nurse: We'll get you some water and take a look.

Patient: I'm not talking about thirst, I'm talking about my Android phone!

Re:How many people...

[nanter]

I spent a couple of weeks in rural Honduras on a medical mission, and we were in a remote village where there was no clean water, goats roamed the dirt "streets," people lived in mud huts, and to get to the closest physician you had to walk about 20 hours to the nearest town. No electricity and no land lines.

Yet, a bunch of people in this town had cell phones, and service was available and reliable. Honduras's cell coverage was far better than the coverage here in the states, and people in dirty poor areas use them as their only means of communication to the outside world, oftentimes to their relatives in the United States. Doctors without Borders put in a clinic that can be used by visiting physicians, and it has solar panels. People pay a nominal fee to charge their phones off the solar panels.

Smart phones? Yeah, I didn't see any of those. But as they become more ubiquitous and less expensive, I can see them taking hold even in places like this.

Re:How many people...

[arielCo]

In GSM the cell size is limited to about 35 km radius, because the frame length limits the amount of timing advance. CDMA is free from such issues, so the max size is largely dependent on frequency, possibly up to 40 mi depending on TRx power and frequency.

Re:How many people...

[e9th]

Where do you live that a 30 mile trip is only a 15 minute drive? Next to an autobahn on-ramp?

Re:How many people...

[Maximum Prophet]

More than you imagine. In Africa, cell phones are the first telephone infrastructure in place, because wiring people's houses is much more difficult than setting up a few towers.

In China, after bottles and such were banned from sports stadiums, people began throwing their cell phones at the officials when they made what the crowd thought was a bad call.

Re:How many people...

[merockstar]

Actually, Verizon covers EVERYWHERE except a small blip in central Ohio that happens to be between two massive hills.

So yea, I know of at least one area that gets great cell service but is still country as all hell.

Re:How many people...

[CharlyFoxtrot]

Cell phones apps don't magically stop working when you have no coverage. And I guess people out in the Australian outback who rely on so-called "flying doctors" (or indeed people relying on similar services around the world) could use this kind of app for a first diagnosis in a pinch. Doubly so if the application could give first aid instruction to people on the scene of an incident.

I dunno...

[mgierhart]

...the whole "dermatologist camera" idea sounds cool, but what about internal afflictions? The article says it can tell you whether you're having heartburn or a heart attack--how? What makes this any more reliable than WebMD? Speaking of WebMD, the self-diagnosis numbers are going to go through the roof. I'm sure the ER docs are just going to LOVE this...

As a non AI physician

[Dunbal]

This is a terrible idea. However I guarantee that the AI algorithm will have a "success" rate of around 85%, since that is the rate at which illnesses spontaneously cure themselves. This rate is why homeopathy, snake oil salesmen, faith healers and all other forms of shamans and charlatans manage to convince people of their effectiveness. Too bad that 15% of the patients will suffer permanent disability or die using these methods. That's the part of the statistic we doctors manage to concentrate on and improve, the 15% that really need help...

Re:As a non AI physician

[BlueBoxSW.com]

That's an interesting stat. I never heard that before, but it makes sense.

As a non-physician, AI hobbyist, this article makes me sad.

Computers have so much potential to improve the quality and decrease the costs of healthcare, yet this almost never happens.

Instead, we end of with complex systems meant to mitigate legal liability, expensive system whose sole purpose is to provide another billable test, and IT departments that care more about building fiefdoms than enabling improvements in patient care.

Want to know the difference between poison ivy and smallpox? You could ask 2 questions to tell the difference, but this wouldn't provide a cool image-recognition demo. No one is going to get an X prize for pushing text to a smartphone.

Re:As a non AI physician

[sznupi]

What if you or your colleagues aren't available (or at least not in sufficient numbers) at a given place? Isn't it worthwhile to at least try to help part of those remaining 15%? (don't tell me you can help all of them...)

BTW, do you have any solid source at hand for thise 85% stat? Might be useful to me.

Re:As a non AI physician

[oddTodd123]

Doctors are notoriously stubborn and arrogant about their abilities, and they refuse to believe that a significant share of medical practice can be routinized to be performed by much lower skilled and educated people. From simple hand washing to using checklists, doctors have steadfastly resisted any change that implies they could be doing their job better, or that someone with less training could do the same job.

Nobody is suggesting the smartphone perform open heart surgery, but if it can use image recognition on a rash to tell you to try calamine lotion before going to see a dermatologist, that can save everyone a lot of time and money. Or, to use a personal example, after I fell on my shoulder, it could guide me through a series of tests (of the type "does it hurt when you do X?") and suggest I may have an AC joint separation and I should see an orthopedist. In the last example, I was originally diagnosed over the phone (by a non-orthopedist doctor) after exactly that experience. The default choice in a case such as mine would be to go to the ER. That would have turned out to be an inefficient and expensive choice and wasted a lot of people's time.

Re:As a non AI physician

[NonSequor]

Any time I see someone wanting to solve a problem of software, I ask myself, "what is the best estimate of the minimum information resources needed to cover all cases of that problem?"

For the problem of diagnosing health problems in humans, I'd say that the current best estimate is the amount of training and ongoing learning needed to maintain a team of human physicians with a diverse range of specialties.

If you want to solve that problem in software, I'd expect you'd have to gather such a team and have them maintain an expert system with their knowledge. So they both have to be specialists in their field and have enough computer knowledge to understand your fancy pants AI system (preferably to the extent that they have at least a vague understanding of its internals). They have to be able to understand what they know, what the machine knows, and how to translate between the two. And if the database accumulates any spurious connections, they have to be able to correct that.

I can't really expect much success from any project with requirements like that. Multispecialists are hard to find, they command large salaries, and at some point they're going to get tired of doing data entry and want to move on at which point you're going to have to find new multispecialists to replace them. At that point you will have to train all over again.

My general feeling is that systems which assist experts are very useful (I could imagine a doctor using a diagnostic database to make sure that alternative possibilities to the most immediate diagnosis aren't overlooked), but systems which attempt to distill experts into software are, at present, doomed to failure.

Re:As a non AI physician

[Dunbal]

BTW, do you have any solid source at hand for thise 85% stat? Might be useful to me.

      Medical school. No seriously, it's in most medical texts that have to do with diagnostics/semiology and/or family medicine texts (Rakel would be a good place to start, I think I remember a break-down in the first few chapters, etc) that break down what patients consult for by disease type. I'm not a walking library and I can't hand you a reference with page numbers, etc, but I'm not making it up. Once you remove the non life-threatening allergies, the common viral infections (respiratory and digestive) that don't become complicated, and all those messy and painful but non life-threatening menstrual problems that women constantly complain about, all the sedentary overweight people with digestive and musculoskeletal complaints, and all the migraines, there's really not a lot left. But that little bit that's left MUST be treated by a physician or the patient will probably die from it.

Re:As a non AI physician

[Dunbal]

Isn't it worthwhile to at least try to help part of those remaining 15%?

      Please explain how your smart-phone will dispense medication and/or apply treatment (assuming it is correct)? Simply knowing what you have will not cure you.

      If you can get to "civilization" where there is a pharmacy, there is also a doctor somewhere close.

Re:As a non AI physician

[Dunbal]

Well I am a physician but I also have a strong programming background.

I used to think that medicine was like a giant algorithm. On one side you had the symptoms. On the other side you had a list of diseases. And in a third column, the treatments for those diseases. I thought that all a doctor had to do was go down the checklist, matching symptoms until they got the right disease, then applying the treatment.

Frankly, this is how medical shows on TV also like to portray it. So non doctors cannot be blamed for thinking that way!

But medicine is much more than that. When I went to med school I found out that the LAST thing we learned was lists. Well that's not true. We had to memorize all sorts of things. But the way we are trained is that we become experts in the human body. First we learn what a normal human body is supposed to look like (anatomy) and how it's supposed to work (physiology). Then we learn about infectious diseases (microbiology) and how they cause the body to break down. Then we learn about other disease processes (pathology) that have their roots in chemical reactions gone awry (biochemistry). THEN we learn how to interview patients (semiology/diagnostics) properly. Then we get put in front of patients for3+ years, seeing people every single day, being quizzed both formally and informally every single day many times a day, to try to get us to think of all the possibilities the guy in front of you could have. After a while you get to learn what's common and what's not.

So although the patient in front of you who doesn't have a pulse COULD have Takayasu's disease, it's far more likely that she doesn't have a pulse because she's dead.

The other advantage that a doctor has over an algorithm is that the doctor WILL be objective. However a patient who is self-diagnosing with a computer will always be subjective. This is of the utmost importance. Doctors do not need to speak to patients. We are perfectly capable of treating unconscious people. This means that when you go to the doctor and ramble on and on about this and that, and you notice he doesn't seem to be paying attention, it's because he's not. What you are saying is no longer relevant. He already has an idea about what is wrong, and is looking for objective evidence. But the subjective patient is likely to diagnose a whole range of illnesses that match his symptoms. The key is to make the CORRECT diagnosis the first time.

Re:As a non AI physician

[sznupi]

At the least it can tell you that you should seek medical assistance, "because..." (instead of relying, say, on some shaman). Nvm that having locally a medical kit with some essetianl drugs is probably more likely than easy access to medical stuff - especially if said drugs would have fairly straightforward dosages, were not really harmful if not needed, but it's not a bad idea to know when & which one will be good for you?

At the least, it will ease the strain on handful of medical personnel that's available / direct cases requiring attention more quickly to them. I think you underestimate the problems present when you have only few phisicians per 100k people.

Re:As a non AI physician

[sznupi]

Note that I wasn't thinking you made it up; it would be just useful to have some reliable and accesible (easily linked to)...something (some knowledge base / whatever). For some time I have my own share of being fed up with how people percieve statistics, most notably in medical matters; this number sounds useful / would be nice to have it at hand.

But I understand how you don't have a need for such kinds of sources; thanks for pointing out the fact itself.

Re:As a non AI physician

[spire3661]

Of course a doctor would say that, god forbid medical answers come for free... People are stupid, but saying that distributing knowledge of this type is a bad idea is just plain shortsightedness.

Re:As a non AI physician

[Thing 1]

[...] "success" rate of around 85%, since that is the rate at which illnesses spontaneously cure themselves. This rate is why homeopathy, snake oil salesmen, faith healers and all other forms of shamans and charlatans manage to convince people of their effectiveness.

So couldn't that same argument be used against doctors?

I mean, if out of 100 diseases that I catch my body will deal with 85 of them, why should I make a doctor's appointment for all 100 of them?

Separately, energy healing works; it took me 40 hours of practice before I could feel the energy (for me, it's a buzzing/tingling sensation; for others it's hot/cold, or pressure, or some other known sense -- it's not a sixth sense, just a reused one). EFT ("tapping psychology") works great as well; just be prepared for some tears (crying, not ripping).

Sure, there are people selling snake oil, but there are also people sharing treatments that have worked for thousands of years. "Western"-style medicine only recently arrived on the scene, and has its own issues with placebos.

You can find value in many places. Of course, when your livelihood depends on selling pills, then you will tend to denigrate the other methodologies whereby pills are not necessary. So I understand where you're coming from, even if I don't want to fill your wallet.

Re:As a non AI physician

[Dunbal]

But I understand how you don't have a need for such kinds of sources;

      I honestly tried googling it for a few minutes, came up blank. I do know I've seen it in at least 2 texts, however, and heard it in several lectures.

Re:As a non AI physician

[Quirkz]

So what you're saying is skip the phone and the doctor, and just wait it out, because the odds are good that you'll get better anyway?

Just kidding, really ...

Re:As a non AI physician

[Anonymous Coward]

Anon because I've modded.
 

The default choice in a case such as mine would be to go to the ER.

Why? I've had numerous orthopedic injuries in my life, and only one of them - the knee that got sliced deep when I fell on a sharp rock - merited a trip to the ER. I've sprained an ankle, torn the medial collateral ligament of my knee, had an AC separation - all of which were taken care of by going to see the orthopedic surgeon within a day or two.
 
The ER is for things that threaten life or limb. Don't go there otherwise. If you break a bone, you probably should go. If the bone sticks through the skin, you definitely must go. If you sprain an ankle? Call the clinic, tell them what happened, ask if you can be worked in today or tomorrow. If not, call another clinic.
 

Doctors are notoriously stubborn and arrogant about their abilities, and they refuse to believe that a significant share of medical practice can be routinized to be performed by much lower skilled and educated people.

The problem with routinizing care to less-educated, less-skilled practictioners is that it very much depends on the less-skilled practitioners being able to recognize when they're out of their depth. This is not as easy as it sounds; viz. the study a few years ago that pointed out that inexperienced people don't even recognize what they're doing wrong - the classic example being sports, where a pro knows immediately after taking a swing that the hit was good, while beginners can't predict which swings will result in grounders to first base and which ones will go over the fence.
 
As for hand-washing and checklists, the underlying issue there is that checklists and rules can and will be abused by bureaucrats in situations where they make no sense. Checklist for central line placement? Great. Checklists when the patient is actively trying to die? Not so good.

Re:As a non AI physician

[Daniel Dvorkin]

the doctor WILL be objective ... Doctors do not need to speak to patients.

Any physician who thinks this way should be required to post his name and specialty, so that prospective patients know to stay far, far away.

Re:As a non AI physician

[Anonymous Coward]

why should I make a doctor's appointment for all 100 of them?

Don't. Also, don't ever ask for antibiotics "just in case". Very few physicians' livelihoods depend on common viral illnesses; in the scheme of things, they won't miss the business.

Re:As a non AI physician

[Anonymous Coward]

Query, if "the AI algorithm will have a "success" rate of around 85%, since that is the rate at which illnesses spontaneously cure themselves" then that should keep the non life threatening people out of doctor offices and hospitals, thereby freeing up space for the people who are near death or mortally wounded, and therefore really need the help.

If one takes an objective view on that 15% who die from using bad methods, then one could hypothesize that they were unwise enough to believe those bad methods, and therefore further the human species by dieing.

Re:As a non AI physician

[BlueBoxSW.com]

While I'm sure that was a pretty honest and accurate account of what it takes to be a doctor, it also highlights some of the reasons for the frustrations people have with them.

Notice how the qualities people say doctors lack fall last in line in the training. You have to think those strong in these areas may have already been knocked out by the 5-10% dropout rate in the early years.

I still think there so much technology could be doing, and that the model of training and interacting with physicians is outdated. The typical $100 for 10 minutes of consultation after which the patient will be referred to a specialist, given an expensive test, a prescription for a drug they can't buy on their own, or told to go home and sleep it off, have to go away.

I know it drives those in the medical field crazy when patients come in with all this information they learned on the internet, but at the same time, thank god for that. One, they are taking responsibility for their health, which should be encouraged, and second, all of the patient education I've ever gotten from a physician has been half-assed. I've always looked up more details and been better educated online (Yes, you have to be careful).

An electronic medical record that spanned your whole life, included your doctor's notes and whatever extra info you wanted to put in (filterable, of course) there would be a great tool for tracking and predicting health over you lifetime, not just a couple dozen 10-minute interactions with a guy that can't remember your name without looking at a chart (no offense, but it's kind of true).

The AI implications would be phenomenal, as well as the simple practical info (want to know what the likelihood of you getting diabetes in the next year is?)

I like what NHS did in England. No one here could post something like this without getting shut down by lawsuits or the AMA.

http://www.nhsdirect.nhs.uk/help/index.aspx

Re:As a non AI physician

[Dunbal]

Any physician who thinks this way should be required to post his name and specialty, so that prospective patients know to stay far, far away.

      You are entitled to jump to any conclusions you wish. I personally enjoy speaking with my patients and fostering a good relationship with them. I have been told face to face and I have "heard" from what people say about me, that I am a good doctor. I never have trouble filling my office. I have never been sued although being human, I have made my share of mistakes. I get stopped in the hallways of hospitals and I do chat with people. I know that some people have named children named after me. HOWEVER the chit chat, the jokes and the HUMAN side has everything to do with inspiring trust and confidence in the patient and nothing to do with reaching a diagnosis. While a clinical history is important, that little pain that your uncle once had years ago is probably not related to what's going on today. But when you walk out of my office you are not going to believe me or follow my recommendations if you think I am a jerk, so usually I end up listening to the full story, including all the things your uncle tried which didn't work.

      Still I find it strange that around a story where people are proposing that doctors be replaced by a cell phone app, you would find objections to my comment.

Re:As a non AI physician

[Dunbal]

then that should keep the non life threatening people out of doctor offices and hospitals, thereby freeing up space for the people who are near death or mortally wounded, and therefore really need the help.

      True. But if it keeps one seriously ill person home because they think they have something less serious and that person dies, then I cannot be for it. While you could argue that the freed up resources will let more seriously ill people be treated and increase their odds of survival, ethically I can't put myself the position of taking such a decision - swapping one life for another.

Re:As a non AI physician

[Dunbal]

because the odds are good that you'll get better anyway?

      Yep. Or look at it another way - if you don't go to the doctor you have a 15% chance of dying. People buy lottery tickets on smaller odds.

      Of course the above is not strictly true. Not going to the doctor will not cause you to have a fatal disease. However IF you have a dangerous disease we can usually prevent you from dying from it. If we can't do that, we can at least provide comfort and relief from symptoms. We can do that for non-fatal diseases too. So all in all it's not a bad idea to consult. All you have to lose is the consultation fee and you potentially have a lot to gain, even if you just have a case of the sniffles.

      Also there is a skew when you look at the distribution. The chances of suddenly developing a fatal disease are small during the middle of life. If you make it to your teens you are likely to make it to your 40's (barring trauma). Death usually happens at life's extremes - the very beginning or the very end. I don't consider 40 to be the end (I'm 42!) but that's when the incidence of heart disease and cancer really jumps. While you can get these diseases at any age, they are not common in say 20 year olds.

Uh oh

[WetCat]

Why do you need AI physician if you already have a phone? Just send a real physician (non-American, of course, because US physicians are really overpriced ones thanks to various professional insurances and malpractice lawsuits) a MMS with your views and symptoms. That will be better analyzed, anyhow.

Re:Uh oh

[ColdWetDog]

As a physician, I've got a short reply for you:

tl;dr

Golf time!

Re:Uh oh

[sznupi]

If you have only few phycisians per 100k people, but where nonetheless there is huge uptake of mobile phones, such AI could surely help in directing scarce resources?

Billions of People.. with smartphones?

[PolyDwarf]

The possibilities are enormous, especially for the billion plus people around the world who live more than a few hours' walk or drive from the nearest doctor."

And yet... they'll have smartphones?

I know, a lot of countries skip land lines and go to cell phones... but all those people who live hours away from the nearest doctor will have smartphones (as opposed to normal cell phones)?

Re:Billions of People.. with smartphones?

[bsDaemon]

In the future all phones will be smart... really, really smart. Maybe they'll even be able to create clean water supplies, build sewage infrastructure, and terraform in-arable land into something productive so that people don't starve, either!

Re:Billions of People.. with smartphones?

[RevWaldo]

In ten years, what we call a 'smartphone' will be a 'normal cell phone'. Think of places like Lagos where in the slum areas you can have access to electricity and cellular phone service but basics like pluming, sewage, and, yes, access to doctors remain a problem.

.

Re:Billions of People.. with smartphones?

[sznupi]

What is a "smartphone" anyway? Latest incarnations of S40 (probably the most popular mobile phone platform on the planet) aren't really that different from the functions, say, iPhone offers; SE "feature phones" even have full multitasking...
And Symbian smartphones are geting closer and closer to the $100 mark (without contract, of course)
The distinction will get blurred / more and more people will be getting "full" smartphones anyway.

There is greater uptake of mobile phones generally than you think; 3.3 billion mobile users at the end of 2007, 4.6 billion at the end of 2009 - now we're getting close to 5 billion probably. Many of them in "developing" places.

Re:Billions of People.. with smartphones?

[Myopic]

People on this thread keep saying that, which is confusing.

To be clear, YES, there are huge numbers of people who have access to cell phones but not doctors.

Re:Billions of People.. with smartphones?

[Anonymous Coward]

yes.

That is the case at least in most of africa, india not so much. There are faciniating socio-encomic drivers to all of this but a bigger percentage of the people in africa have a cell phone than in europe.

The poorest areas might have just one smartphone in the village but then that should be enough for this kind of thing.

Hell it would be nice if my phone could give me instructions on how to deal with someone else in a medical emergency.

Re:Billions of People.. with smartphones?

[Lord Ender]

Yes, they do have smartphones. In India old smartphones are common. Remember that iPaq you got rid of to get your BlackBerry? You can't imagine anyone wanting to use such an old turd of a WinMo phone... but these things get recycled and resold in the third world. Public places there are like museums of old ringtones.

AI automobile driver

[stony3k]

I guess this prize is already won, then. See http://www.automobile.com/stanford-and-vw-self-driven-vehicles.html

Re:LOL no Tricorder tag?

[RangerRick98]

My first thought was "Please state the nature of the medical emergency."

The AMA is gonna be pissed

[Anonymous Coward]

Considering how unfriendly the American Medical Association originally was to giving Nurse Practitioners diagnostic/prescriptive authority, I can't imagine they're gonna be happy about conferring it onto inanimate objects.

Re:The AMA is gonna be pissed

[pseudorand]

> inanimate objects.

Except for the iPhone, which doesn't run Flash, most smart phones can do animation these days.

dumb smartphone idea

[jackspenn]

Especially for the billion plus people around the world who live more than a few hours walk or drive from the nearest doctor

This is brilliant logic, because all of the peole I know who live more then a few hours from the nearest doctor are all about 3.5 hours from decent 3G coverage (more if they have AT&T).

People are retarded, what's next we plan on giving kids who are primarily worried about starving or being shot at by the rival drug gang in Africa a laptop?

Re:dumb smartphone idea

[Krneki]

Especially for the billion plus people around the world who live more than a few hours walk or drive from the nearest doctor

This is brilliant logic, because all of the peole I know who live more then a few hours from the nearest doctor are all about 3.5 hours from decent 3G coverage (more if they have AT&T). People are retarded, what's next we plan on giving kids who are primarily worried about starving or being shot at by the rival drug gang in Africa a laptop?

Of course, this way we only need a couple of laptops with the disclaimer "upon death the laptop returns to the previous owner".

Re:dumb smartphone idea

[Myopic]

You must know people in Illinois or something. If you knew people in rural Africa and India, you'd have a different impression. In many places, cell phones are the only widespread technology, and access to phones far exceeds access to medicine.

Re:dumb smartphone idea

[sznupi]

This isn't about country in which AT&T operates; there are more. You seem to have some weird image of those, too.

Mobile drug test

[desertjedi85]

Immediately proceeding this app will be the mobile drug test app. Just pee on your phone and see what color it turns.

Re:Mobile drug test

[Spazntwich]

I peed on mine and it went off.

Strangely, my girlfriend had the same reaction.

Lesson learned: Waterproof your sex bots.

Take two asprin

[pseudorand]

print "Take two asprin and call me in the morning";

Where do I pick up my check?

Words fail me

[Anonymous Coward]

I have lived in places and seen people who are few hours walk/drive away from a doctor. Let me tell you, if you are few hours walk away from a doctor, then you probably do not have an internet connection let alone mobile internet. A doctor would come to your town far ahead of the Internet facility.

Diagnostics versus differential diagnosis

[debrain]

The notion of creating a diagnostic heuristic isn't particularly novel, and given the frailty of human memory and other inadequacies of the human brain it has the potential to eliminate a significant number of errors in medical diagnoses. The problem with a machine doing this, in my experience, is the diagnostic component -- many diagnoses are based upon the physician's qualitative interpretation of the patient's symptoms (i.e. as experienced diagnosticians).

Nevertheless, somebody clever will undoubtedly eventually create a medical diagnostic tool on a smart phone, which will be of incredible value – reducing the cost of both physicians and their diagnostic mistakes in the medical services industry. Let's just hope such an application doesn't become another pharmaceutical delivery medium (as I have heard physicians referred to by some marketing people in big pharma).

Re:Diagnostics versus differential diagnosis

[jedidiah]

IOW: a machine will be as good as a "nurse practicioner" or a mediocre physician.

I HATE it when someone wants to shove a nurse-pretending-to-be-a-doctor in my face. Mediocre HMO doctors are bad enough.

Re:LOL no Tricorder tag?

[Yvan256]

And after reading your post, my first thought was "I'm a doctor, not a doorstop."

The Simpsons

[DaMattster]

This brings back memories of The Simpsons when Lisa types in Homer and Bart's symptoms into the virtual doctor. The virtual doctor a la Steven Case spews out, "You've got .... leprosy"

Reality imitates Science Fiction

[Anonymous Coward]

" Please state the nature of the medical emergency "

Re:LOL no Tricorder tag?

[Robotech_Master]

Mine too. So naturally I had to search the comment threads to see if someone had posted it already. :)

Thank You For Calling iDoc

[smitty777]

# Press 1 for heart attacks
# Press 2 for aneurysm symptoms
# Press 3 for severed limbs or profuse bleeding

Please stay on hold while we contact a physician for you

[..musak...]

Re:LOL no Tricorder tag?

[Anonymous Coward]

It's okay. I'm sure the developers will get around to fixing your personality subroutines eventually so that you have a more varied introductory phrase.

Great Idea

[Anonymous Coward]

True, but doctors these days are often in a hurry and don't do a great exam either (and charge a bunch of money). I'd rather have the info available to me, and use a doctor as a second opinion if I feel it warrants it. Ultimately, my health is my responsibility.

That reminds me of...

[Anonymous Coward]

a escape pod episode on a handheld boyfriend

http://escapepod.org/2009/09/17/ep216-%CE%B2oyfriend/

I'm a doctor, not a smartphone

[PJ6]

With voice recognition we can use Voyager's Doctor as the screen UI. Maybe use the built-in camera so he can actually take a look, too. I say it's a bit early for this, but we're close.

Facebook status update

[hamburgler007]

has herpes.

Re:Facebook status update

[Anonymous Coward]

Ah, but there's a treatment for that. *

* May not cure herpes

A job for an expert system

[seibai]

It's been relatively well established that expert systems can have high success rates for diagnosing diseases within a limited scope(for instance, http://en.wikipedia.org/wiki/Mycin).

I can imagine something more general could be put together. The main cost is in acquiring the data; the coding itself is trivial. And you could have it tell you to see a human physician when the confidence level is low. A heavy disclaimer would probably be necessary though.

Re:LOL no Tricorder tag?

[Runaway1956]

I anticipate a lot of penis shots. Especially after reading yesterday's slashdot story - http://tech.slashdot.org/story/10/06/15/1419224/Chatroulette-Working-On-Genital-Recognition-Algorithm

Location, location, location...

[Lemental]

What if the rash is in an area that is a little more personal than the leg? How can we integrate this with Chat Roulette?

For the rest of us

[Anonymous Coward]

As an American (estadounidense) living in a very small rural town in Mexico, where what often passes for a doctor is the guy who but a sign reading 'Doctor' out side of his house. And a medical exam consists of poking your belly and prescribing a wide spectrum antibiotic, I don't think this is an all together bad idea. Yes, for those fortunate enough to live near actual medical professionals I can easily see how it would be very problematic and perhaps lead to more harm than good.

It fills the same role that books like "Where There Is No Doctor: A Village Health Care Handbook."

Did the phone pass?

[Krau Ming]

I will only allow my phone to be my doctor when it passes the requisite exams (ie: USMLE, MCCEE).

Re:Did the phone pass?

[Kazymyr]

Having actually taken those exams, I can tell you that there is very little correlation between one's ability to pass them and the ability to practice as a physician.

Portable testers are more useful

[Animats]

It's a cute idea, but probably works about as well as self-diagnosis through Google.

More useful would be portable, automatic test equipment for blood samples. This is already available (PowerPoint file, will open with Open Office Impress) in a small desktop machine. It's marketed, though, as a device which allows US medical practitioners to do blood tests in their offices while charging insurance companies the usual price an outside lab would charge. These things need to come down to the cost of a high-end smartphone, be deployed in pharmacies, and operated by people at the nurse/practitioner, paramedic, or pharmacist level.

Can I borrow your phone?

[Comboman]

Can I borrow your phone? I need to do a colonoscopy.

Expect a lot of push-back on this.

[maillemaker]

I have seen this coming for some time, as I'm sure many people have.

Most of the time when you go to the doctor, he is guessing at your symptoms based on a verbal patient interview. Unless the symptoms have a physical manifestation that the doctor can see or measure, diagnosis is largely guesswork.

There is no reason why this sort of guesswork needs to be done by a highly trained and expensive person. The only reason why we do is we are afraid of the consequences of guessing wrong so we feel better by having they highly trained expert do the guessing.

But we are entering an era where automated diagnostic tools, such as the camera tool in the article, are going to become a reality. We are not yet to the "Star Trek Tricorder", but I don't think it will be long and you /could/ go to a kiosk at the drug store and be scanned, including a blood analysis, by an automated piece of machinery that will be able to diagnose based on observable information rather than guessing based on how the patient is able to describe their symptoms.

This is going to eat into doctor revenues massively.

For society, this will be a good thing. Doctors will be freed up from diagnosing ailments to concentrate on actually curing ailments.

But doctors are not likely to take kindly to these new tools. Just like today there is pushback against the mall-store ultrasound clinics that people can use to get ultrasound images of their unborn babies. The doctors are claiming that these services are "unsafe", despite the fact that you don't have to be a rocket scientist to operate the things. They just don't want their business undercut.

Phone - could it be Lupus?

[91degrees]

Seriously, half a dozen Voyager EMH jokes and not a single House reference?

Sure it can!

[denzacar]

Rectal exam physical enough for you?

DAMMIT JIM!

[denzacar]

It was set on "car" not "cancer"!

Also... obligatory...

[denzacar]

I'm a general practitioner not an oncologist!

Odd prioritization

[Daetrin]

The addendum just mentions in passing this is only one of four new X Prizes, however the other three all sound much more interesting than an AI Physician for cellphones. Organ regrowth/replacement and AI cars could have a huge impact on our lives, and even a new deep sea submersible sounds kind of cool. As others have pointed out there are already a lot of health resources on the web that are accessible by smartphone. This seems more like evolutionary development than revolutionary.

As a physician myself...

[flipper9]

As a physician myself (well I'll have my MD next year), there is more to a diagnosis than a simple test/analysis. The most important part of any diagnosis is the "History of Present Illness" as well as the physical exam. Knowing when to use certain tests, and how to interpret them in the light of the particular patient's present condition is what makes medicine an art as well as a science. I think the idea of a camera taking a picture of a dermatologic problem being a great tool, but it won't be able to make a conclusive diagnosis. Many tests are not black and white, yes or no answers to a problem...they themselves contain errors, which if interpreted incorrectly can lead to further harm to the patient. I think it's going to be awhile before we have an AI doctor like "The Doctor" from the Star Trek Voyager series.

Cheaper and easier than effective AI

[Beorytis]

Why not just send the photo to a call center in Asia that's staffed with actual physicians?

Similar to telehealth

[Landreville]

In Canada (at least Yukon and Ontario) there is a (free of course) phone number where you can chat with a nurse any time you want. They generally tell you what you can do or whether you should go to the emergency room. It's super handy for people that live outside of cities where it is a long drive or possible even a plane trip to the nearest doctor. Implementing a smartphone application where instead of just consulting WebMD, it actually connects you to a nurse would be the best route to take with this. It would be similar to a phone call, but you could include photographs, or even audio recordings of how you scream when you touch the bubon on your leg.

Holographic doctor?

[Anonymous Coward]

I hope it has a better bedside manner than the one from Voyager.

Iphone Blood Sugar App

[wsanders]

http://www.diabetesmine.com/2009/03/lifescans-new-diabetes-iphone-app.html

This is one step. This thing could save lives - unless it's Flash-based.

/. next poll

[dugenou]

Slashdot's next poll: how many hours' walk are you from the nearest doctor?

They don't need AI for this

[surfcow]

As I read it, they aren't looking for AI, they are looking for a expert system with a decent interface. The ability to weasel data out of the patient and interpret it. A big database of facts to bang it against. There is no reason it would have to have a voice interface - text would do fine, better initially.

People throw the term AI around a lot.

Re:They don't need AI for this

[Lord Ender]

Expert systems are classified as AI.

Brilliant demographical decision

[ThatsNotPudding]

The service requires a smartphone - which in turn requires a hefty chunk of steady income for the initial contract and the $30 per month data plan (and the mandatory credit card). And who can afford such a costly, high-tech gizmo? Those who already have health insurance. "What, you're sick? Serves you right for being poor!"

Virtual Doctor ala Simpsons

[noncommercial]

"You have Leprosy. Good bye."
*Virtual Doctor avatar scurries away*

on photographing a rash

[Harvey Manfrenjenson]

Having your AI analyze a cell phone photograph of a rash isn't likely to help in most cases-- you need to know if the rash is palpable, if it blanches when you press on it, exactly how it is distributed over the skin, etc. And a cell phone photograph of smallpox lesions would be completely indistinguishable from shingles.

Lousy example, in other words.

Fool

[Anonymous Coward]

Yeah, just like Professional Engineers "protect" their "monopoly" on being certified to build structures like bridges and buildings competently so that people's lives are not unnecessarily endangered.

Get a freakin' clue.

Must be...

[Anonymous Coward]

Must be a brain aneurysm.

http://www.webmd.com/brain/tc/brain-aneurysm-topic-overview

MEDgle - an on-demand accessible medical AI

[ashdamle]

Towards this and to assist patients,nurses,physicians, hospitals, clinics, etc we developed MEDgle http://www.medgle.com/ . We're still in beta but feedback, suggestions and more are very welcome. We believe strongly in empowering everyone with the best personalized evidence based options anywhere anytime. :) Cheers Ash

Processing...

[Markos]

(AOL VOICE)

You've got.... AIDS.

So, just press a button on your phone....

[mrpolyrhythm]

and a friendly face pops up saying: "Please state the nature of the medical emergency." Bonus if your phone has a projector!

Replace doctors!?

[bussdriver]

There are plenty of cut-and-dry situations where computers can replace or out perform humans and even when they are sub-par, billions of people around the world have limited to no access to doctors but increasingly are getting smart phones. Some people have no medical but they have a cell phone!

This is just the beginning; better hardware and years of development in this area will produce machines that can do even better. It should someday come down to cheap doctor free centers or drug store machines to do the bulk of the diagnosis work before referral to a specialist. A phone will always be limited because it will not have the hardware to probe - a digital x-ray booth with recognition software could make that whole area accessible to billions who couldn't have it before. There is no way a human doctor can know enough to diagnose it all (they luck out in that it follows a bell curve;) the subject is too vast -- somebody has to be on the front line and it may as well be a cell phone that helps you find the proper specialists.

Errors will happen but I see no reason why machines can't out perform the humans on average. I've seen plenty of doctor errors in my life...

This challenge is easy.

[cbreaker]

Simply make the thing say "Could be cancer" for every type of ailment.

I win!

It's cancer!

[Aereus]

Remember -- it's always cancer.

Add it into Chatroulette

[Culture20]

Want to know if that rash on your leg is poison ivy or smallpox? Take a photo of it with your phone and ask the AI.

Or add this into chatroulette and inform the trollers what STDs they have.

This would be amazing...

[crhylove]

... for myself and for millions of hard working Americans without health care. I can afford a phone bill once a month, but I certainly cannot afford to do anything but die in a hospital.

Really, AI on a smartphone is the best solution?

[haggais]

If I already have my smartphone with me, or indeed any phone with camera, and I can take pictures of my skin condition (or whatever), and write or talk about it, perhaps I could SEND the photos and comments to a REAL LIVE doctor, even one who works more than a few hours drive away, but within phone coverage, and get his NATURAL INTELLIGENCE comments, instructions, etc. over the phone? Is AI going to be better and cheaper than human doctors any time soon? And assuming it does get that way, why does it need to run on my smartphone? Through the magic that is the interweb (of which I believe smartphones are a part), would it not be easier to send the data to a slightly more powerful SERVER, and get my AI GP's advice from there? Slapping "AI" and "smartphone" on a problem does not make for a brilliant futuristic solution.

Re:Really, AI on a smartphone is the best solution

[DriedClexler]

What doctors would actually provide this service?

Oblig. Idiocracy

[Chess Piece Face]

"This one goes in your ear, this one goes in your mouth, and this one goes in your butt. No, wait..."

Artificial Dr. House

[kauttapiste]

Ooh ooh, I want the Dr. House edition!

"Tests take time. Treatment's quicker."

Settle A Bet

[JustABlitheringIdiot]

Boil or Pimple? Oh wait it's a gummy bear. Mmmmm.... Gummy Bear....

Perfect face-tracking cameras first

[s_p_oneil]

http://consumerist.com/2009/12/hp-face-tracking-camera-doesnt-track-black-faces.html

If we're having a hard time getting relatively simple AI image recognition tasks like this to work, the ability to make an AI doctor for a cell phone is a very very very long way away.