pageok
pageok
pageok
Brain Scanner Evidence Used for Conviction in India:

The International Herald Tribune reports:

[W]ell before any consensus on the technology's readiness, India has become the first country to convict someone of a crime relying on evidence from this controversial machine: a brain scanner that produces images of the human mind in action and is said to reveal signs that a suspect remembers details of the crime in question....

[I]n June, in a murder case in Pune, in Maharashtra State, ... a judge explicitly cited a scan as proof that the suspect's brain held "experiential knowledge" about the crime that only the killer could possess, sentencing her to life in prison.

Psychologists and neuroscientists in the United States, which has been at the forefront of brain-based lie detection, variously called India's application of the technology to legal cases "fascinating," "ridiculous," "chilling" and "unconscionable." ...

"I find this both interesting and disturbing," Henry Greely, a bioethicist at Stanford Law School, said of the Indian verdict. "We keep looking for a magic, technological solution to lie detection. Maybe we'll have it someday, but we need to demand the highest standards of proof before we ruin people's lives based on its application." ...

"Technologies which are neither seriously peer-reviewed nor independently replicated are not, in my opinion, credible," said [J. Peter] Rosenfeld, a psychologist and neuroscientist at Northwestern University and one of the early developers of electroencephalogram-based lie detection[, speaking of the particular approach used in the Indian case -EV]. "The fact that an advanced and sophisticated democratic society such as India would actually convict persons based on an unproven technology is even more incredible." ...

Thanks to GeekPress for the pointer.

PersonFromPorlock:
Interesting. What changes in prosecution and defense would reliable lie detection technology entail?
9.16.2008 2:00pm
ejo:
until our criminal justice system is completely free of any taint, we should hold off on comment. after all, we should be concerned about keeping our own house in order, not worrying about possible injustice in India.
9.16.2008 2:05pm
hattio1:
Experiential knowledge of the crime? So, she could have witnessed it right?
9.16.2008 2:07pm
Crunchy Frog:
What did the minority report say?
9.16.2008 2:11pm
Hoosier:
"What's it going to be then, Alex?"
9.16.2008 2:12pm
Norman Bates (mail):
A bad solution to an enormous problem. My observations lead me to conclude that many of the very worst people are excellent liars and some of the best do not come across well in front of strangers. In US trials this problem is compounded by the fact that rules of evidence often prevent juries from getting the very information that would allow them to appropriately judge a witness's veracity.
9.16.2008 2:15pm
Angus:
The consequences of having a truly accurate machine that will catch 100% of people who break the law are so horrifying the public will never stand for it. Pretty much everyone out of childhood is guilty of at least one legal infraction for which they have not been caught, even if it is something as small as speeding or jaywalking.
9.16.2008 2:15pm
C Miller (mail) (www):
Well, here in the U.S. a Seattle scientist claims that he has developed an electronic brain test, which uses a technique called "brain fingerprinting" and which could be more reliable than the polygraph test. Meanwhile, some researchers are claiming that fMRIs are producing 100% accurate results. It looks like we might be entering a brave new world. Will the one-eyed man be king?
9.16.2008 2:27pm
Cindy:

Experiential knowledge of the crime? So, she could have witnessed it right?



According to the full article:

The inventors of the technology claim the system can distinguish between peoples' memories of events they witnessed and between deeds they committed.
9.16.2008 2:33pm
Jon Roland (mail) (www):
This points to a general problem of scientific laypersons, such as most judges and jurors, trying to reach judicial findings based on "expert" scientific witness testimony. Most people are not educated to have the kind of skepticism most good scientist have, and tend to project their hopes and fears into technology they don't understand.

What is usually missing from such testimony is a good faith estimate of the level of confidence that should be placed on it. Witnesses are browbeat into testifying "are you sure? Yes or no."

As an eyewitness to a street crime myself, who picked the culprit out of a lineup, I am sensitive to the problem. I never had to testify, the accused plea bargained, but if I had I would have had to testify in the form, "The probability is less than 5% that the accused is indistinguishable from someone else of the same ethnicity and less than 10 years different in age." As a trained observer, for me that would be a good estimate, I used to measure it by testing my reliability in recognizing distant aircraft by make and model while I was an Air Force Air Traffic Control Officer. Most people don't get such training, and therefore make poor witnesses, resulting in a high rate of convictions of innocent persons by eyewitness testimony that may be sincere but is not reliable.

What was missing from the testimony in this case was a record of reliability of identification of guilty persons using the method, with appropriate double-blind controls, and evidence for the lack of pressure on the witness to report a positive. (A problem with the FBI Crime Lab revealed by the whistleblower Frederick Whitehurst.)
9.16.2008 2:34pm
Bored Lawyer:

The consequences of having a truly accurate machine that will catch 100% of people who break the law are so horrifying the public will never stand for it. Pretty much everyone out of childhood is guilty of at least one legal infraction for which they have not been caught, even if it is something as small as speeding or jaywalking



Why should this bother anyone. So a machine can prove I committed a petty infraction at some point in my life -- I once jaywalked or double-parked. So what? These petty crimes carry no moral stain, and at worst a minor fine (assuming you are caught and prosecuted soon enough.)
9.16.2008 2:36pm
theobromophile (www):
Isn't this just the latest and most high-tech version of hair analysis?

Silly question, but would this be permissible in the US? Would there be a Fifth Amendment claim that would prevent its use, or are your brain waves not considered to be "testimony," much in the way that your handwriting and voice are not considered to be so?
9.16.2008 2:36pm
some dude:
Self-incrimination taken to the highest degree.
9.16.2008 2:38pm
C Miller (mail) (www):
PersonFromPorlock, I'm not sure whether this directly answers your question, but, in New Mexico, polygraph test results are presumptively admissible. Maybe somebody has done an empirical study of how this has changed the behavior of prosecutors, defense counsel, and the courts. If not, maybe I will conduct one.
9.16.2008 2:38pm
Houston Lawyer:
Computers are magical. They can even tell you that man-made carbon emissions are causing the earth to get warmer.

GIGO
9.16.2008 2:39pm
Curt Fischer:
What Jon Roland said, especially this part, which in my mind bears repeating.


This points to a general problem of scientific laypersons, such as most judges and jurors, trying to reach judicial findings based on "expert" scientific witness testimony. Most people are not educated to have the kind of skepticism most good scientist have, and tend to project their hopes and fears into technology they don't understand.

What is usually missing from such testimony is a good faith estimate of the level of confidence that should be placed on it. Witnesses are browbeat into testifying "are you sure? Yes or no."

9.16.2008 2:47pm
Gregory Conen (mail):
@C Miller: Or not. You have to be careful about claims made by people who have a vested interest in the success of lie detection machines.

Most (all?) of the studies done on novel methods are in low-stress, laboratory conditions (high-stress situations like criminal trials may cause false positives) and with inexperienced liars (sociopaths, etc, may be able to fool the new devices, like they can polygraphs).

Maybe this is the new DNA (and even DNA is not so as straight-forward as it may seem (do a search for DNA testing "birthday paradox"), but the verdict is not in yet.
9.16.2008 2:50pm
C Miller (mail) (www):
Gregory, I agree with all of your points. I raise these example with a high degree of skepticism and the fear that these tests might find more success than the polygraph test.
9.16.2008 2:53pm
Malvolio:
until our criminal justice system is completely free of any taint, we should hold off on comment. after all, we should be concerned about keeping our own house in order, not worrying about possible injustice in India.
Uh, why? Yes, our system is imperfect -- every system is imperfect -- but if someone else is committing a great injustice, we are obligated to point it out.

That said, I don't see why this case is particularly unjust. Fingerprinting isn't perfectly accurate, nor is DNA testing, nor is any other technology. That doesn't mean technology should be banned from the courtroom (in favor of what? Eyewitness testimony?)
Pretty much everyone out of childhood is guilty of at least one legal infraction for which they have not been caught, even if it is something as small as speeding or jaywalking.
Don't you think that's a problem with the legislative process, producing laws that the majority won't obey?
9.16.2008 2:55pm
Roger Schlafly (www):
This is not really lie detection. The technology has nothing to do with what the perp says. It merely detects whether he has seen the scene before.
9.16.2008 2:56pm
Ben P (mail):
But if I had I would have had to testify in the form, "The probability is less than 5% that the accused is indistinguishable from someone else of the same ethnicity and less than 10 years different in age."


A HA!!

So there's a *Chance* you were wrong, so was it him or wasn't it?

I think that's also very much the same reason many trial lawyers attempt to avoid having Jurors that are likely to be extremely analytical during deliberations unless it carries a very obvious advantage to their case for some factual reason.

It may even have something to say about that study that showed that a significant percentage of self identifying democrats want judges to "be fair" in their decisions. To a great many people who aren't conditioned into thinking a certain way a decision that holds something obviously "unfair" because of a substantive rule of law or a procedural technicality just doesn't compute.
9.16.2008 2:57pm
David Schwartz (mail):
First of all, this type of application is way outside the wildest claims for what this technology may ultimately be able to do. What it may be able to do is distinguish between forming memories and recalling memories.

Thus you could show someone a picture of a crime scene and get insight into whether they had seen that scene before. Perhaps you could construct several different reconstructions of a crime scene (with different body positions, say) and see which one invokes a recall response.

That might be possible in the not-too-distant future. But using it as a lie detector or distinguishing between witnessing and doing -- that's not only not currently possible but not likely to be possible.

As for whether such evidence should be admissible, you can think about similar hypotheticals. What if there was a machine that could flawlessly read out your memories of where you were and what you were doing in a particular time frame.

Frankly, I think we've already slid too far with handwriting exemplars.
9.16.2008 3:04pm
Ben P (mail):

This is not really lie detection. The technology has nothing to do with what the perp says. It merely detects whether he has seen the scene before.


The interesting part of this, if I'm recalling the same research I read on this sort of thing before is that in certain circumstances, (hallucination or dreaming for example) the brain activity looks almost exactly identical to someone who actually remembers the secene.

The only real advantage I see in this particular technology is that while it's not necessarily mo re reliable than a polygraph, it's harder to spoof. A person who's simply has their lie down so completely that they won't show any nervous indicators on a polygraph can't necessarily control their brain patterns as easily, unless of course they happen to be a buddhist monk or something.
9.16.2008 3:05pm
David Schwartz (mail):
This is not really lie detection. The technology has nothing to do with what the perp says. It merely detects whether he has seen the scene before.
You are confusing the legitimate experimental version of this technology (which tests whether one is making a memory of a scene or recalling one) with the nonsense version used in this case. Nobody claimed to be testing whether anyone had seen anything before in this case. In fact, nothing was shown visually to the subject -- her eyes were closed.
9.16.2008 3:06pm
Scote (mail):

some dude:
Self-incrimination taken to the highest degree.


No, it is much, much worse than that. It is alleged self-incrimination given an artificial scientific imprimatur of 100% accuracy.

You are guilty if they say you are guilty, even though there isn't a single peer-reviewed study on the technology, no proven error rate. It is a prosecutor's wet dream come to life. And it is an abomination of the judicial process.
9.16.2008 3:11pm
spring (mail):
The best market for this device will be the self-help, supplement-taking consumer. Millions of these individuals will pay $1000 flex-pay to find out what they are thinking. They can self-diagnose mental illness in the privacy of their own homes and then pursue a brain-feedback therapy. Sort of like Scientology's "tin-cans."

I think this technology should not be used in a legal setting because it probably would not be able to distinguish between the witness's brain waves and those of the demon that is in possession of it
9.16.2008 3:29pm
Opher Banarie (mail) (www):
PersonFromPorlock: What changes in prosecution and defense would reliable lie detection technology entail?
Well, after applying it to the defendant it could be used on defense counsel!
9.16.2008 3:48pm
argh:
Jesus, when Ron Paul was talking about all this happening months ago, I thought he was crazy. That crazy old fart was right all along.
9.16.2008 3:49pm
argh:
Oops, meant to post that previous comment in the "good news, bad news" thread. Mea culpa.
9.16.2008 3:50pm
Thief (mail) (www):

What does a scanner see? I mean, really see? Into the head? Down into the heart? Does a passive infrared scanner ... see into me - into us - clearly or darkly? I hope it does see clearly, because I can't any longer these days see into myself. I see only murk. Murk outside; murk inside. I hope, for everyone's sake, the scanners do better. Because if the scanner sees only darkly, the way I myself do, then we are cursed, cursed again and like we have been continually, and we'll wind up dead this way, knowing very little and getting that little fragment wrong too.
9.16.2008 3:53pm
wfjag:
Ain't science wonderful!

"Eating veggies shrinks the brain", Times of India, 14 Sep 2008 on http://timesofindia.indiatimes.com The story reports that a scientific study found, among other things:


Scientists have discovered that going veggie could be bad for your brain-with those on a meat-free diet six times more likely to suffer brain shrinkage.
[and]
Brain scans of more than 1,800 people found that people who downed 14 drinks or more a week had 1.6% more brain shrinkage than teetotallers. Women in their seventies were the most at risk.

Beer does less damage than wine according to a study in Alcohol and Alcoholism.

So, hand me a brew and tell where that damn cow is. I think it's time to exercise a 2d Amendment right.
9.16.2008 4:08pm
Mikeyes (mail):
This kind of surface EEG technology has been around for years and all sorts of claims have been made for it with very little convincing proof that it can do anything more than record electrical waves from the surface of the brain. When compared to depth electrodes, surface measurements are not accurate at all.

EEGs can detect overall changes, can pinpoint (with only general accuracy) electrical changes that are fairly crude such as a seizure locus, and have a fairly good record in helping guide a neurologist in finding brain problems.

Detecting lies is not one of the things that is well documented. Since the precision and accuracy of medical EEGs is not all that great and the fact that there are plenty of false negatives (and positives, but negatives are more likely) I am not sure how any reputable scientist can support the claims made by the purveyors of this machine.

Early in my career I was involved in computer analysis of surface EEGs, depth electrodes and voice in a project looking at these very questions. While the work did produce some interesting results (almost all of it not published due to the funding, you can figure that one out on your own) none were what I would call so accurate that they could be used in a court room. In fact, after an initial trial with these methods in the real world, they were discarded as not being practical enough for the lettered organization who hired us.

I wish I had the cites to show you, but they don't exist. I would like to see the science behind the claims being made in India.

BTW, what were the alternatives offered this woman? She had to have undergone the testing voluntarily.
9.16.2008 4:13pm
Randy R. (mail):
This is off point, but interesting (at least to me).

A friend of mine is a renowned linguist, and she did ground breaking work on lie detectors and language. Her research showed that we all have an original language, the first one we learned. But then sometimes people will immigrate to another country and learn that language. They may no longer use their first language, or use it only sporadically. Or they may use it all the time.

However, if you subject them to a lie detector, it is much easier to successfully lie in the second language than in the first. So her conclusion was that for criminals whose first language is not English, you should do the test in the first language.

There are many reasons for this, of course, but it shows the limits of lie detectors, and how you really have to know a lot about them to use them correctly.
9.16.2008 4:25pm
John Moore (www):
Assume the technology is valid and verified.

Would compelling someone to take such a test be a Fifth Amendment violation?

Wouldn't allowing the refusal to take such a test into evidence also be a violation?
9.17.2008 1:24am