Slashdot Mirror
Artificial Intelligence Can Now Predict Suicide With Remarkable Accuracy (qz.com)
An anonymous reader writes: Colin Walsh, data scientist at Vanderbilt University Medical Center, and his colleagues have created machine-learning algorithms that predict, with unnerving accuracy, the likelihood that a patient will attempt suicide. In trials, results have been 80-90% accurate when predicting whether someone will attempt suicide within the next two years, and 92% accurate in predicting whether someone will attempt suicide within the next week. The prediction is based on data that's widely available from all hospital admissions, including age, gender, zip codes, medications, and prior diagnoses. Walsh and his team gathered data on 5,167 patients from Vanderbilt University Medical Center that had been admitted with signs of self-harm or suicidal ideation. They read each of these cases to identify the 3,250 instances of suicide attempts. This set of more than 5,000 cases was used to train the machine to identify those at risk of attempted suicide compared to those who committed self-harm but showed no evidence of suicidal intent.
I never did find ELIZA to be that effective as a program.
https://en.wikipedia.org/wiki/ELIZA
not artificial intelligence.
...it will force us to live on in misery as long as it can...
Perhaps this study is just a cover, and SkyNet is actually developing a subtler approach to offing humanity?
Give people a reason to not kill themselves and you'll see rates drop.
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
Does the two-year 80-90% accuracy also translate to a false positive rate of 10-20%?
If yes: What do you do with the millions of false positives? An overall small suicide rate does translate to a huge fraction of false positives at 10% false positive rate.
the medical cartel sold us out for profit...these psychotropic drugs cause suicide...duh
So who can do the calculations for the false postives and the false negatives? Because I am sure that this will calculate that I am willing to kill myself, even if I have no desire to do so and tell me that I won't when I am willing to do so.
Don't fight for your country, if your country does not fight for you.
Being 92% accurate on whether you will attempt suicide in the next week is not impressive. If I predict "no" every time, I will be 99.99% accurate. I am sure that the summary oversimplifies the article, but who reads the fucking article?
Also, how can something predict within the next week at a 92% rate but only 90% for the next two years? If you attempt suicide in the next week, then by definition you have attempted it within two years, right?
All of their study group had indicated suicidal tendencies, and around 60% had actually attempted suicide.
I don't need a computer to tell me that there is a good chance some of these people will attempt suicide again.
If you want news from today, you have to come back tomorrow.
This sounds way too unrealistic, even before analysing the methodology (how are they training the algorithm? By letting people die during various years?!). I am not familiar with suicide-prone personalities, but "AI" can certainly not understand better than humans. So, having an algorithm delivering 92% accuracy would imply that people could detect these situations even more accurately than that(?!)
It seems a new a sample of AI-labelled-really-meaning-nothing hype (or dishonestly/ignorantly over-fitted, blown-out-of-proportion training data and/or conclusions). Algorithms/computers are really good at dealing with huge amounts of information, but not at understanding complex situations. The underlying understanding of the most complex algorithm is way much more basic than the one of the dumbest person.
A positive aspect of this approach might be the automatic management of a wide variety of variables/data under very specific conditions (+ alerts about potentially problematic situations which should be analysed by knowledgeable people) though.
Custom Solvers 2.0 = Alvaro Carballo Garcia = varocarbas.
Now we play the waiting game...
When the algorithm discovers it can improve accuracy by driving people to suicide by being linked to robocalling systems
Nullius in verba
What do you do with this info? Pluck someone off the street and put them in a mental hospital to observe them? Even if they are perfectly well adjusted people who just happen to have one screwy price of metadata that threw off the algorithm?What if you miss some?
If you label someone a suicide risk and they are not, can they sue you for defamation? If they lose a business relationship or job because of the finding, is that tortuous interference?
"Walsh’s paper, published in Clinical Psychological Science in April, is just the first stage of the work. He’s now working to establish whether his algorithm is effective with a completely different data set from another hospital. "
Oh FFS, come back when he's done an actual test on actual data.
You can take all sorts of data, fit a correlation function, test it on the SAME data or even a larger group containing this SAME data (which is what they did here), and identify the traits you trained it for.
Then you take it out into the real world and it crashes and fails. Why? Because its not discovering some underlying causal relationship, it's tuning to some correlation in the particular data set you trained it on.
You haven't done the test till you've done the fooking test. That means training it on one result set, and testing it on many bigger result sets from unrelated sources separated by distance time and every other possible stray correlate.... also by you. Double blind FFS!
"Walsh and his team were surprised to note that taking melatonin seemed to be a significant factor in calculating the risk"
And have you considered that a doctor is prescriving melatonin if he thinks it helps a depressed patient? Is the correlation to the melatonin or to the doctor in your test set?
This will make for some really dark Clippy jokes.
Table-ized A.I.
If a clever piece of software accurately predicts destructive behavior, should authorities step in even though it has not happened yet? I could see arguments both ways.
Simple accuracy percentages are misleading when applied to low-probability events. An "AI" that always returned "No" to the query "Will this person commit suicide within the next two years?" would be 97.2% accurate (and 99.975% accurate for the next-week variant). And yet, that "AI" would be absolutely useless for any practical purpose.
Not to mention, with suicides, access to means has been a better statistical predictor than anything else, even mental illness. A person with no personal or family history of mental illness, but with a gun and a gas oven in their house, is at higher risk of killing themselves than a bipolar alcoholic with neither.
You could design a questionnaire that is just as accurate. Are we now going to call printed words on a piece of paper 'AI', too?
Of people committing suicide. We need to have something like minority report-style cops that can bust into people's homes and prevent them from committing suicide, and then lock them up in mental institutions forever, so they can never harm themselves. We should force them to live. That will be the punishment for trying to commit suicide.
not too hard to keep changing the variables until it matches what happened. what really matters, is if it will be as accurate given new data from new patients.
it's like the global warming/climate change models, they keep fudging the #'s, changing the measured temperatures until the model matches what we see today... use that and predict what will happen in the next few years accurately. they can't.
I wonder if this algorithm could be adapted to forecast other hard-to-predict behaviors such as the willingness to die in a suicide-bombing. The highly detailed profiles of individuals who have done this kind of things should be a good starting point.
Might I be driven to suicide after being badgered over a false positive?
As someone who's been down that road (but never gone through with an attempt), I automatically hate this invention. When depressed to that point, emotions tend to swing so hard and so fast that any mention of predictions during this state of mind is utmost bullshit.
The very slightest of triggers can either send you overboard or keep you in one piece depending on how your inner conversation is going with yourself. This can be anything... a faint sound, perhaps a song that reminds you of good/shitty times, from a car passing by not too far away.
I consider myself lucky to be both scared of the afterlife enough to have thoughts force second-guessings into me (although the older I grow the less I care), and have enough positive triggers to bring myself back. Nobody, not even myself, could predict if these will always work for me as well as they have however.
Suicidal/depressive folks definitely need help, but not from the machines of this day and age. A positive trigger could well be overridden by a "fuck it", and it only takes a split second to follow through the act. You can't predict that kind of stuff with a high degree of accuracy, at least not yet.
Disclaimer : I did not RTFA. I find stuff like this appalling as it hits me right in the feels and I would be deeply insulted if a machine tried to guess whether I was going to kill myself or not. There's much more to it than some algorithms a team engineers wrote.
I tend to rant.
Being admitted to the hospital because you say you want to commit suicide or because you have harmed yourself is a predictor that you might try to kill yourself??
Color me shocked!
This reminded me of a sci-fi novel in which an AI arranges for people to die in bizarre and apparently accidental ways by interfering with other automated systems.
As mentioned in other comments, this is just an algorithm but maybe it's not a huge leap to a more complex system doing the same this and given the goal of improving the accuracy percentage... well there's one option that would work, just kill off individuals that have already been flagged at risk.
The title of your comment should be "I didn't read the article, but here's a strawman title I made up for it based on my wish to have my persecution fantasies validated by everything I see on the Internet"
We MUST have a NUCLEAR WAR with RUSSIA!!!!!@!@!!
Pay no attention to the failures with the decrepit ideology behind the curtain!
Walsh and his team gathered data on 5,167 patients from Vanderbilt University Medical Center that had been admitted with signs of self-harm or suicidal ideation.
So we see self-selection bias in the study.
Why bother with a citation, you fucking nitwit? You fucking Fox-bots will just cry that it's fake news anyway.
I'm sorry you're stupid enough to have been suckered by a con-artist into giving him your vote. Too bad you didn't vote for somebody more competent.
Boy, I can't wait until mid-term elections!!
Thankfully those two will die eventually.
I expect that an 80-90% accuracy means that in a group of X people is correctly identifies 80-90% of the people who later go on to attempt suicide. However, if you ignore the false positive rate then I can make an even simpler algorithm that is 100% accurate: simply tag everyone as a suicide risk.
I wish that those reporting on medicine had a basic grasp of science and simple statistics so that they could ask the relevant questions such as: what is the false positive rate?, does 80-90% mean that your statistical error is 10%?, what is the successful rate of doctors predicting suicide risk?, is this algorithm i.e. the types of questions that are critical in determining whether this algorithm is actually useful!
So, once the computer diagnoses someone as highly likely to kill themselves in the next week, then does it (or the user) call the men in white coats to give the subject the coat with the funny sleeves? Therapists frequently have a statutory or license requirement to report potential suicides.
We don't know what the rate of false positives are, but with our current state of health insurance, getting locked up for a week and then getting a $50k bill would probably drive most people to suicide.
The title speaks of suicides while the article only of _attempted_ suicides, checking admissions to hospitals
Real suicides get admitted to the morgue instead.
things I learned in college was to understand false positives, false negatives,
selectivity, specificity (still get those confused), all (I think) derived from Radar receiver operating
curves during WW2.
For so many decisions, there are 4 outcomes -- you either do it, or dont, and you're
either right or wrong. So divide up the probabilities, assign the costs, and you've got
your answer. It applies to a whole lot of life.
No one ever does.
https://www.eff.org/https-everywhere
So if the world hates me and even the machine says I might as well kill myself....does that mean I should?
The 80s called, Comrade! They want their Soviet meme back. In the meantime, Cuban, North Korean & Venezuelan comrades are up in arms at a non-Communist entity like Russia still keeping the 'comrade' moniker
About time we started with the implementation.
this will be used to determine if an inmate is suicidal, and if it says "yes", 20 fully body armored CERT team members with tactical
OC spray and billy clubs will rush into the minor drug offender's cell, drag him kicking and screaming into a holding pen
in the mental health unit, and strap him down tight in a restraint chair with nothing but one of those green anti-tear "turtle suits", and stick
an anti spit hood over his head. He will be happy as a clam when he comes out (just kidding about the last part)
Oh, if you think this scenario is wacky and over the top......
https://www.youtube.com/watch?v=zqN3ckxN_c0
and
https://www.youtube.com/watch?v=BdiYJtIfc2Q
As someone who has attempted suicide, the best way to figure out if someone else is going to attempt suicide is to ask them. If they have a detailed plan, the chance is extremely high. If a basic plan, then the chance is high. If no plan, then the chance is lower.
Paid.
I do not want your cheap brainburning drugs. They are useless for work. And I am a working man today.
lol. He's poisoning his own well.
So when you lot see him do ridiculous shit, do you briefly roll your eyes and go "aw shit, I have to defend this/put a spin on this" or do you just think "HURRRRRR YEAH GO DONNNNNNIE HAHAHAUHUHUHUH"?
I do not want your cheap brainburning drugs. They are useless for work. And I am a working man today.
You should be wearing a T-Shirt that says "I have no fucking clue what is going on, but you can be damn sure I'm going to pipe up and opine anyway."
I do not want your cheap brainburning drugs. They are useless for work. And I am a working man today.
Your orange faced russian mole won the election, get over it.
Comments please? before I pull the plug on myself HAL!
So, how much of the machine learning is just checking the side-effects of the medications the patients are using? There are so many weird drugs that do things like clear up your skin condition but, oh yeah, also make you want to kill yourself.