Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
Technology

A Utah Company Says It's Revolutionized Truth-telling Technology. Experts Are Highly Skeptical. (washingtonpost.com) 51

Is the ocular product EyeDetect a leap ahead of the polygraph? Or just the same dubiousness in a more high-tech box? From a report: In 2018, John Rael, a volunteer track coach in Taos, N.M., was on trial for allegedly raping a 14-year-old girl when his lawyer made an unusual request. He wanted the judge to admit evidence from "EyeDetect," a lie-detection test based on eye movements that Rael had passed. The judge agreed, and five of the 12 jurors wound up voting not to convict. A mistrial was declared. EyeDetect is the product of the Utah company Converus. "Imagine if you could exonerate the innocent and identify the liars ... just by looking into their eyes," the company's YouTube channel promises. "Well, now you can!" Its chief executive, Todd Mickelsen, says they've built a better truth-detection mousetrap; he believes eye movements reflect their bearer far better than the much older and mostly discredited polygraph. Its popularity may be growing: the company says EyeDetect has gone from 500 customers in 2019 to 600 now.

Its critics, however, say the EyeDetect is just the polygraph in more algorithmic clothing. The machine is fundamentally unable to deliver on its claims, they argue, because human truth-telling is too subtle for any data set. And they worry that relying on it can lead to tragic outcomes, like punishing the innocent or providing a cloak for the guilty. EyeDetect raises a question that draws all the way back to the Garden of Eden: Are humans so wired to tell the truth we'll give ourselves away when we don't? And, to a more 21st-century query: Can modern technology come up with the tools to detect those tells? An EyeDetect test has a subject placed in front of a monitor with a digital camera and, as with the polygraph, lobbed generically true-false queries like "have you ever hurt anybody" to establish a baseline. Then come specific questions. If the subject's physical responses are more demonstrative there, they are presumed to be lying; less demonstrative, they're telling the truth.

This discussion has been archived. No new comments can be posted.

A Utah Company Says It's Revolutionized Truth-telling Technology. Experts Are Highly Skeptical.

Comments Filter:
  • by nospam007 ( 722110 ) * on Wednesday November 17, 2021 @03:50PM (#61997091)

    I can see it in their eyes.

  • by Chelloveck ( 14643 ) on Wednesday November 17, 2021 @03:51PM (#61997095)
    If something like this actually worked, a measly 20% increase in customers in two years is absolutely pitiful. Seems about right for snake oil with a charismatic salesman, though.
  • by laie_techie ( 883464 ) on Wednesday November 17, 2021 @03:52PM (#61997103)

    Remember that lying is defined as telling something one knows to be false. Also, this technology measures how nervous / anxious the person is. Merely being accused of a heinous crime is enough to shake people so they fail the test despite being innocent.

    • The real issue is how do you design a control for this. I can imagine that a lie where there is an expectation you are lying (like as part of a study) will have a different response than lying under psychological duress (ie questioning by police or in court).
      • The real issue is how do you design a control for this. I can imagine that a lie where there is an expectation you are lying (like as part of a study) will have a different response than lying under psychological duress (ie questioning by police or in court).

        How do you have a control for someone who has been shat on unfairly by a corrupt system that is then questioned by that same system and whose answer will subject them to imprisonment? You can’t, any reasonable human would suddenly get nervous when they know they are being royally fucked over.

    • Remember that lying is defined as telling something one knows to be false.

      But what does "knows to be false" mean? A sociopath can convince themselves that almost anything is true, if it benefits them. "I won the election!".

      • Remember that lying is defined as telling something one knows to be false.

        But what does "knows to be false" mean? A sociopath can convince themselves that almost anything is true, if it benefits them. "I won the election!".

        Yes, a sociopath can tell a lie while manipulating their body to not show any nervousness, or anything else picked up by polygraphs.

    • The idea of discovering an absolute reality, i.e. truth, from the recollections of biased observers by a program trained by biased observers is cute.
  • by jellomizer ( 103300 ) on Wednesday November 17, 2021 @03:58PM (#61997119)

    We believe in our lies all the time.
    Did you rape that person? Well no. (because you convinced yourself that your victim, was sending you signals that they wanted it)
    Did this person rape you? Yes (because you have convinced yourself that you are the victim, because in hindsight you realized the consequences)
    Did you provoke the violence? No (because you were just goofing around, the other person took it out of hand)

    This is the reason why, the justice system appears to be sloppy, lie detectors don't work, because we believe in our lies, by not forcing ourselves to overthink the situation, and just stay in your frame of mind at the time.

    • by gweihir ( 88907 ) on Wednesday November 17, 2021 @04:13PM (#61997177)

      Exactly. This tech cannot work. Self-delusion is one of the most evolved human mental capabilities.

      • by Ă…ke Malmgren ( 3402337 ) on Wednesday November 17, 2021 @05:15PM (#61997349)
        No, it can't be! I refuse to believe... oh.
      • It's not just self-delusion.

        People nowadays are also not self aware and do not have any idea for to empathize. I see people complain about other people cutting them off when they are driving (after they just cut someone off), leave empty food boxes in the cabinets but getting mad when others do that, etc.

        I have watched my wife's kids get mad when other people's trash is on the floor. When I ask if they mean the fast-food bag that's been on the floor for a week, they say "No - it's the napkin next to it.

      • It could definitely work, during an interrogation. After the fact it's as useful as the interrogator saying "In my expert opinion, he's telling the truth". Neat... but a jury shouldn't hear that crap.

        Doesn't mean it doesn't work. Normally you'd go by your gut if you think someone was bullshitting you, and circle back to those topics and try to trip them up. A machine could definitely be a, well a _different_ set of biases for sure, but another signal to hey, let's ask them that again a little differentl

    • Lie detectors don't work because they are looking for involuntary responses people are said to make when they lie that, if they exist at all, are buried beneath noise.

  • proved you can get away with fake products for quite awhile profitably. I guess we will find out if Lizzy pays any price for it soon.
    • proved you can get away with fake products for quite awhile profitably. I guess we will find out if Lizzy pays any price for it soon.

      What do you mean fake?!? Why, with just one drop of an eye, we can get at a whole ocean of truth! Believe me!

  • by MacMann ( 7518492 ) on Wednesday November 17, 2021 @04:05PM (#61997151)

    For lie detection to work the person has to know they are lying. They also have to feel some kind of shame, fear, or other emotion for this to show up in their eye movement or such. It's a fairly common trope used in movies for people to pass these tests through deception, training, or some hand waving science or magic. This is a trope in movies because it's often discussed in real life.

    One example of a failure, or excessive difficulty anyway, was in Blade Runner. The person under evaluation was given false memories to cover up the truth. In another movie the people could swear they saw an alien spaceship take their friend away, because in their minds at least that is what they saw. Is that what happened? We don't know, all we know is that they believed this as true. Lie detection doesn't tell you the truth, it tells you what people believe to be true. Given the wide margin for interpretation in this system mentioned in the fine article I suspect this is just as flawed as any other system for lie detection.

  • Psychopaths believe their lies so could probably pass this easily.

  • Unless they have a bunch of data from scans done on people convicted of rape who were convicted without using the scan as evidence I don't understand how this can even get off the ground. It does not make sense to me that a rapist doesn't have differences in their mental makeup that bleed into their other behaviours (like dishonesty) and render the analysis extremely suspect.

    The model has to be built on a database of known innocent and guilty people, and all of them have to believe the questioning is real,

    • Someone mod Petersko up; he understands the real problem: lack of data to evaluate (and possibly lack of even more data to train, if it's based on machine learning, rather than "Golly, I have a hunch").

  • I don't think you need technology to tell the truth. Phrasing does matter...
  • Use it on them first, Bond-villain style.

  • by SkonkersBeDonkers ( 6780818 ) on Wednesday November 17, 2021 @04:17PM (#61997199)

    In all lie detection techniques is they are tainted by the interrogator and since interrogators are almost always profit making companies/individuals, they have a vested interest in giving their "customers" the results they want. So if hired by law enforcement, they favor a lie outcome. If hired by the defense, they favor a truth outcome. Because you know what happens when you don't give customers what they want, they don't hire you anymore.

    Even polygraphs themselves in theory are extremely accurate. In practice though, they are terrible.

    Also not that long ago the rage was micro-expression reading and it turned out that had the same fundamental problems without the woo factor of a graphing machine and electronic sensors to give even a semblance of objectivity since it was just entirely the subjective viewpoint of some supposedly highly-trained interrogator who would wave their hands a lot and give a lot of impressive sounding technobabble but really it comes down to whatever they want to say biases and all.

    • by Ichijo ( 607641 )

      Even polygraphs themselves in theory are extremely accurate. In practice though, they are terrible.

      You misunderstand. The purpose of polygraphs is not to determine whether someone is lying, but to make them think you know they're lying so they confess.

      I thought everyone knew this but I guess not.

    • "Even polygraphs themselves in theory are extremely accurate. "

      Citation needed. All blind study I saw about polygraph is that they barely kick above the 50% value, they only work well on stuff the examiner is already aware of the answer on and where the examined is not aware they are quack, and a lot of innocent people fail them due to nervosity (after all polygraph only measure your heart/blood pressure lung reaction, and so they measure nervosity and NOT lying), and are highly subjective from the exami
      • no, they're refused as evidence because they infringe the fifth amendment/privilege against self-incrimination/right to silence.

        now it also turns out that polygraphs are pseudoscience, so that's another reason to not use them.

  • by sjames ( 1099 )

    People with autism tend to have atypical eye movements, in particularly gazing appropriately at others while speaking. I have to wonder if they have controlled for this in any way. There are many other conditions that may cause unusual eye movements.

    Either way, without a metric assload more information, this looks like more snake oil right up there with phrenology and the polygraph. I would hope our courts have learned their lesson, but I'm not so sure they have.

  • between false negatives and false positives, you will be just as effective at telling if a random person is lying by flipping a coin.
  • Revolutionized Truth-telling Technology

    A better wrench [xkcd.com]?

    • by vivian ( 156520 )

      A better wrench?

      Only if it delivers a hefty electric shock every time it detects a lie.

      What I want to know is how did the subject respond when asked about the good things that come into his mind when thinking about his mother?

  • True sociopaths do not believe they have done anything wrong, do not believe that deceiving or manipulating someone else is wrong as long as they get what they want, will believe their own version of events over all other views, even if caught on video, and could kill someone and pass a polygraph test while still covered with blood.

    Eye movements caused by "guilt" or "remorse" will not happen with the people who are truly a danger to the rest of society.

    relevant trivia; there are people who can consciously c

  • ... eye movements reflect their bearer ...

    Does the subject believe they committed they committed a crime? No.

    Does the interrogator manipulate the thoughts and emotions of the subject? Yes.

    Is a 'skilled' interrogator using only "prove your innocence" paranoia and hypnosis? Yes.

    Does the subject feel guilty or anxious? Yes

    All truth-detectors; polygraph, micro-expression/micro-tremble, fMRI; assume a specific physiological response means manipulation of their words, a re-writing of their thoughts. That's a problem because brains are very i

  • ... generally being able to identify when a person chosen at random is lying is isomorphic to solving the Halting Problem [wikipedia.org]
  • by King_TJ ( 85913 ) on Wednesday November 17, 2021 @06:10PM (#61997503) Journal

    I've seen some studies where they apparently had some level of accuracy at gauging who is lying and who isn't by studying eye movements. But these were mostly conducted with people speaking on television. It seems like that's a very different scenario, right off the bat, because the individuals on TV had no idea someone was going to try to analyze them for truthfulness.

    If you're getting interrogated as part of an arrest, you're going to be in a much different frame of mind answering their questions and having conversation with them....

  • I, and most people on the spectrum, hate eye contact. How are you going to calibrate for someone lying vs just on the spectrum? The behaviors vary quite a bit. I know many special-ed kids, including my own and their friends...it's not easy figuring out if they're lying. Various cultures discourage eye contact. How do you account for someone growing up on the South Side of Chicago vs a rural village in China?

    I am skeptical these folks have done enough research to make this admissible in a court of l
  • For most interesting questions, the person being questioned may have convinced themselves of something that is factually incorrect.

    Its very difficult to do a fair test for use in trials because you would need to create a very high stakes situation with strong emotional involvement. The person being questioned would need to be in real fear of negative consequences. Otherwise you might just create a "stress test" rather than a "lie test".
  • One of the things Five Eyes do is align their security clearance standards. Makes sense for information sharing right ?

    The US is the only one that ( for some fucked up reason ) uses polygraph in their vetting process.
    Everyone else ( rightly ) considers polygraph to be bullshit snake oil. If its inadmissible in court due to unreliability then it shouldn't be trusted anywhere.

    This eye tracking thing is more of the same.

    Anyway, if you ever want to defeat a stupid polygraph, put a stone in your shoe and press d

  • They just need to invent the memory recall device, and they will have a Za`tarc detector: https://stargate.fandom.com/wi... [fandom.com] lol

    Suffice it to say, this has bullshit and abuse written all over it.

  • Describe in single words only the good things that come into your mind about ... your mother.

I judge a religion as being good or bad based on whether its adherents become better people as a result of practicing it. - Joe Mullally, computer salesman

Working...