Slashdot Mirror
Brain Scans to Identify Liars?
dotc writes "After a bunch of sci-fi stories and rumors, now it looks like the future has become a reality -- a reliable, unbiased test using functional MRI brain scan to detect lying. The article author details a first-person account of undergoing the MRI 'deception task'. And the test is available now - use it to prove your innocence." From the article: "Laken said he's aiming to offer the fMRI service for use in situations like libel, slander and fraud where it's one person's word against another, and perhaps in employee screening by government agencies. Attorneys suggest it would be more useful in civil than most criminal cases, he said."
But advocates for fMRI say it has the potential to be more accurate, because it zeros in on the source of lying, the brain, rather than using indirect measures
This is completely bogus. Look, if one can lie (and is good at it), it is going to be much more difficult to figure out whether they are telling the truth or not. To someone who knows what they are doing, polygraphs can be fooled and I would suspect that interpretation of fMRIs can also be confused by someone who "knows" how to lie. The trick is to avoid delivering "tells" that are physiologic manifestations of deception. The truth is that there is no foundation in physiology that mandates that one has to reveal anything when stating something that is not in fact, the truth. A good liar will be able to deceive the device and more importantly, the interpreter of the device because they are able to LIVE the lie.
Now, I am not saying that all means of determining lies by technology are doomed to fail. Rather, I believe that relying on any one (particularly trendy) method for determining lies will work. And the use of fMRI is simply a massively expensive and trendy polygraph, particularly because there are so many differences in cortical anatomy and regional differences between individuals. I would be much more comfortable with a derivative of cortical function such as the p300 cortical recognition waveform used as part of a more complete determination of truth using interview, cross checking of facts, polygraph and p300. Perhaps if the fMRI proves accurate to some degree, it could be integrated, but it should not be used exclusively.
And yes, I do know a little something about neurophysiologic monitoring as I teach neurophysiology labs to medical students.
Visit Jonesblog and say hello.
"It's not a lie, if you believe it."
What's the MRI gonna tell you then?
Lie detectors have always been more of a psychological test than an actual method of detecting lies. That's why they're not admissible in court, nor can an employer force you to take one. Now suddenly they can read your brain patterns (which they don't actually understand, just generalize) and tell if you're lying?
I don't buy it. I'll believe that they have a more accurate method of telling when you experience psychological stress from lying, but the actual act of lying is such an indistinct thing that I can't believe that you have a portion of your brain that says "turn this on when you lie".
The fact that they want to make this admissable in a court of law is just plain scary.
Javascript + Nintendo DSi = DSiCade
This won't work for those who have mental issues and actually believe they are telling the truth. When they scan your brain all the 'sectors' will still show up as true. However, this would still be useful after it has undergone some extended testing to ensure accuracy.
as detecting truth.
What's more, they admit it doesn't actually detect lies, because people beat it; and that's under idealized lab conditions.
Do not go directly to jail.
KFG
The for profit lab reports the test is accurate 90 percent of the time. Even after an independent study is performed, I'm still not sure I'd trust the accuracy. Controlled tests (where subjects are directed to steal an object) are very different than real world scenarios. Regardless, I suspect that, like polygraph tests, courts will eventually rule the outcome of such a procedure is not admissable evidence.
An effective signature identifies a particular user amongst a base of thousands.
If the subject truly believes the response to a question regardless of it's validity, there's much you can do in the way of physical monitoring.
If the subject is telling you what they believe to be true, then they aren't lying. They may be incorrect, but that's not the same thing. This device is useful for detecting when someone is knowingly giving untrue responses. Seems to me it would be highly useful. I'd like to see the Enron execs hooked up to this thing for a little Q&A.
It's not enough to bash in heads, you've got to bash in minds. - Captain Hammer
I'd like to point to: http://www.abc.net.au/science/k2/moments/s1213245. htm
It's an article talking about how easy it is to implant memories that never existed into peoples minds. In fact, not only do people end up remembering things they've never seen, but they also end up adding additional information to the stories. It's a bit scary actually, but it's a good thought on how one might "break" the system.
Quoting the article:
"It's one thing when implanting false memories is a laboratory experiment, but it's quite another when the accused wrongly end up in jail..."
Anyone remember the time when you were considered innocent until proven guilty?
Last summer, there was as a piece on This American Life about a man going through a lie detector test as part of the process of obtaining a security clearance. Everything went fine until they started asking about child pornography. The guy freely admitted that he looked at porn, but he conceded that at some point, inadvertently during one of his porn viewing sessions there might have been an under aged person in one of the pictures. He didn't know for sure either way, but since he suspected that it was probable that in all the pornographic pictures he's seen an underage person was present at some point, he couldn't answer the question "Have you ever looked at child pornography?" with a definite "no", and in the end received no clearance and had answered questions in such a way that made him out to be a pedophile, despite the fact that the worst thing he did was look at porn too much.
The problem with a purely 100% accurate 'truth telling' system is that it's too easy to neglect to measure intent or look at grey areas, especially when one freely admits to a minor infringement of the law or policy which put them inadvertently in a worst position. For example, in my younger days, from 1992 to 1996 I used to smoke marijuana on a pretty regular basis. I don't think it's a bad thing, and even though I don't do it anymore (I just don't feel like it) I have no issues with telling anyone who asks about it. Despite it being against the law, I don't see it any more dangerous then exceeding the speed limit by 10 mph or jaywalking.
However, nearly 7 years ago I returned to my apartment one night from a particularly difficult day at work. One of my neighbors offered me a pipe of what I assumed was marijuana, which I accepted. I took a long draw on it, and noticed it didn't taste anything like what I was used to (and for that matter, didn't look right burning in the bowl). I said to my neighbor 'this is some really weird weed', to which he replied 'It's not weed, it's crack'. I don't even know if it got me 'high', I was so pissed off. I spent the next 4 or 5 hours in a fit of rage walking around the block. I never spoke to that neighbor again. To me, this was a big deal.
And now, if somebody put me through any 'truth' machine, and asked me about drug use, I'd have to say that I have, in fact, smoked crack. A device like this combined with specifically directed questions could easily paint me as a real junky, even though I'm not and I have some pretty strong feelings about the harder, more dangerous drugs like cocaine or heroin, and even though I haven't smoked (nor have desired to smoke) marijuana in over two years.
I would hope, though suspect that it won't come to pass, that certain measures would be put in place that would look at intent or degree before reaching a conclusion. If I was asked 'have you ever stolen anything', the answer would be yes. 26 years ago, when I was 5 years old, I took a matchbox car from a local supermarket without paying for it. I still feel guilty about it, and haven't stolen anything since. If absolutes were used and I was obliged to be completely honest, I'd end up being thief in addition to being a junky.
No man or woman is compleatly without sin, and without looking at intent a machine like this could be used to make anyone look like a monster.
The Internet is generally stupid