LAFAYETTE POLYGRAPH NO LONGER SELLING CONVERUS EYEDETECT

[Ex Member]

"...despite the uncertain theoretical underpinning, CQT's seem to be reasonably accurate."

Dr. Vrij's comment is correct, but "reasonably" is a subjective term. If you read deeper you'll see that the CQT can have a low specificity which makes it biased against the innocent. "Reasonably accurate" could mean better than chance, a notion which I tend to agree with. Slightly better than chance could be appropriate as part of an aggregate risk calculation. But to view the results as deterministic is what generates so much displeasure from posters in this forum.

Perhaps the launch was not premature from a business standpoint which has other criteria like seizing a niche, heading off competition etc. My call of a premature launch resulted from noticing the scant field testing thus far conducted. I also do not feel at ease with what they used to establish ground truth.

Regarding customers, you will rarely find a purchaser of a new cool Camaro complain--they are still intoxicated with their new toy.

[Neal Harris]

Ark:  Your are absolutely correct!   The CQT is biased against the innocent, which is why we publish the FN and FP error rates openly.   I don't agree that we view the results as deterministic.  We instruct our customers to do what Wandersmann suggests:  interrogate, deep background check, and exonerate when there is no supporting evidence of deception.  Our customers clearly understand EyeDetect's error rates, and that they MUST be taken into account before using the tool to ruin someone's life.

Dr. Kircher would love more field data, but it is very hard to come by.  We have tried and tried.  It took years to get the field studies approved.  In the US, it is virtually impossible for University researchers to test on human subjects.  Nevertheless, we are continuing to push for more.  Could anyone on this thread share a good polygraph screening field study? 

Ark, which specific element of ground truth was concerning to you (hair, saliva, or confessions)?  We know there are error rates in each, but if you have any ideas for our scientists on how to better establish ground truth, I will forward to the science team.

Regarding customers, please re-read my post on stickiness.  We have many customers that order monthly.  We also have government customers running thousands of tests per year.  In Mexico, one casino chain ran 2,600 tests last year and just reordered for the next 12 months.  For them, this is not a cool new toy.  They have integrated EyeDetect into their HR software and workflows using our API's (application programming interfaces).   They feel EyeDetect is more fair, less biased, and more scientific than polygraph examiners using hand scoring.

[Ex Member]

They feel EyeDetect is more fair, less biased, and more scientific than polygraph examiners using hand scoring.
                   

This is marketing pablum. Prove it.

I will respond to your other points later. I have to catch an international flight.

[Tom Tesslin]

I do not understand how my initial post about Lafayette dropping Converus EyeDetect ends up in numerous posts that really have nothing to do with the core values of this site.  If Converus is so sure of what they have to sell and the government (Whose Govt?) is buying so many licenses (whatever those cost?) why is this Harris asking all of these questions?  He appears to be a salesman that is 100% sure of his product, so why seek out information from others on this site, other than to use it for his advantage?  In going back through previous posts no other lie detection or credibility assessment salesman has been posting about his or her product and/or seeking solutions, ideas or other data to improve the product. If I had any knowledge that could improve that device or any other lie detection device I certainly would not provide it. All of these devices can be manipulated by the user/examiner to obtain the answers he or she, or his agency ultimately is seeking.

[Ex Member]

He appears to be a salesman that is 100% sure of his product, so why seek out information from others on this site, other than to use it for his advantage?  In going back through previous posts no other lie detection or credibility assessment salesman has been posting about his or her product and/or seeking solutions, ideas or other data to improve the product.

Hmm, you have a point Tom. Neal, let me know if Converus would like to negotiate a consultation contract.

[Neal Harris]

Neal, let's put "business decisions" aside.  Do you yourself believe that EyeDetect would be about 86% accurate in specific-issue fidelity cases?  Also, does Converus regard polygraph as science, or as pseudo-science?

Note - I need to figure out how to put a box around a the text box, so I will highlight for now.

Dan and Tom:  I would need to consult with Dr. Kircher on the likely accuracy for a fidelity test.  We have not discussed it to date.  An R2 would need to be created that had face validity, which is usually the challenge in a CQT as Ark highlighted.

Tom, why do you care which tests we decide to develop?  When you work at a tech startup, you place bets on certain markets.  We started in LatAm and then US federal.  We then moved to many other countries via Lafayette's international partners that signed partnership agreements with us.  We are also pursuing PCSOT and US law enforcement.  We have plenty on our plate without aggressively pursing the fidelity market.

Science vs. Pseudoscience response is still being drafted....

[Neal Harris]

Hmm, you have a point Tom. Neal, let me know if Converus would like to negotiate a consultation contract.

Ark:  Absolutely!  We have consulting contracts with media relations consultants, lobbyists, business services consultants (myCFO, etc.) and technology consultants (typically ex-researchers from US federal agencies).  We also have a technical advisory board.  From your posts I believe you could make a significant contribution.  What are your rates?

[Neal Harris]

He appears to be a salesman that is 100% sure of his product, so why seek out information from others on this site, other than to use it for his advantage?  In going back through previous posts no other lie detection or credibility assessment salesman has been posting about his or her product and/or seeking solutions, ideas or other data to improve the product.

Tom, go back to the chart showing science vs. pseudoscience.  Scientists welcome criticism.  It's called peer-review.  Dr. Kircher sent his work to Ben Shakar at Hebrew University, one of his biggest critics but a very respected scientist.  Ben's biggest worry was about EyeDetect being used covertly, but Dr. Kircher doesn't think this will ever be possible.  Sensor technology would have to come a long, long way.

I have made it clear that EyeDetect is not perfect, and has error rates that we publish openly.  Why wouldn't we want to improve the science by seeking input for critics?  Just because the polygraph and CVSA folks don't post on this site seeking input doesn't mean it's a bad idea. 

[Neal Harris]

If I had any knowledge that could improve that device or any other lie detection device I certainly would not provide it.

Tom, why not?  Don't you want to help people that are being victimized by polygraph?  If you could make it better, you wouldn't do it?  That does not seem to align with the core values of antipolygraph.com.  If you made an alternative to polygraph better, maybe it would replace the technology you despise.

All of these devices can be manipulated by the user/examiner to obtain the answers he or she, or his agency ultimately is seeking.

EXACTLY!  The human manipulation is the primary flaw in the process.  The instrument just gathers physiological data.  The humans at the controls interpret and manipulate the data to suit their objectives and agency goals.    For this reason, computer algorithms score and deliver the results of EyeDetect tests, not humans.

[Neal Harris]

This is marketing pablum. Prove it.

Ark:  Why is it hard for you to believe that customers would like an alternative to polygraph?  If polygraph is as bad as everyone on this site discusses, why wouldn't customers like something different?

For proof, go to our YouTube channel.  We are adding customer videos every week.  See the link:  https://www.youtube.com/watch?v=caqz-GH18-o&t=245s

[Tom Tesslin]

I have no opinion on what test you develop. The potential for abuse is there with EyeDetect, as once you sell these units what does Converus care what is done with them or who suffers from improper usage.  My comment was regarding sales volume in terms of the fidelity testing as Dan Mangan brought up a good point. That is easy money. Selling to the USG is no easy task and you have an army of polygraph examiners and related people that will fight you in every possible way. While you may not want to accept it or realize it, Lafayette dropping EyeDetect was not helpful as it gave you instant credibility of sorts in a very specialized market. Personally if I was Lafayette management I would have kept the device, made a few sales every quarter and strung Converus along as to see how the unit progressed. They are probably now in rapid development now of a eye sensor that they can integrate into what they have presently. I bet you the customers are demanding it! They already have a MIND PING device they are selling now. Even if the sensor only works to some degree, it would discourage the customer base from defecting and augmenting with EyeDetect.  You can be sure that ALL  polygraph manufacturers, schools, and other interested parties are going to end up in the same room coming up with legitimate ways to discredit or eliminate EyeDetect. Let us not forget the attorneys that are still crying over Clinton losing the election, as they might end up suing Conversus on behalf of some immigrant or some applicant that failed.  That is what attorneys do, regardless of them being right. Keep you friends close, your enemies closer. Lafayette forgot that old adage. With no disrespect intended, you are in sales, you have nothing to lose for the most part. If things do not work out at Converus, you can  jump on to the next craze in technology and show the same enthusiasm as you have demonstrated here.

[Neal Harris]

I do not think that too many examiners would want his or her data stored on the cloud. It is more secure left on the Lafayette or Limestone polygraph where it cannot be compromised.

Tom:  You are correct.  To address customer security concerns, we have the ability to redact all PII (personally identifiable information) in the data file being uploaded, rendering it worthless to a hacker.  The customer has a key to associate the test results with an individual.  All that is uploaded is the 1's and 0's of the sensor data, where they are scored by our algorithms on our servers.  Also, we use 256 bit encryption and two-factor authentication to access the test results (Google authenticator, Microsoft authenticator, etc.).

Some world governments have laws that restrict any data from traversing their borders (i.e. Ecuador and Colombia), so we also developed a "local scoring" option.  They can run and score tests locally without being connected to the internet.

[Neal Harris]

The potential for abuse is there with EyeDetect, as once you sell these units what does Converus care what is done with them or who suffers from improper usage.

Tom:  EXACTLY!   I doubt you'll believe me, but we do care.  I also believe that the instrument manufacturers care when their products are used to abuse others.  You would have to be a pretty cold and heartless human being to not care about the abuse.

For this reason, we do not allow examiners to create their own tests.  This limits EyeDetect's utility today to our standardized tests, and certainly hurts sales.

We could easily open up our test creation portal to them, but we have low confidence in their ability to adhere to the strict criteria we have designed for a "proper test".  EyeDetect is "locked down" and tests are only created by Converus and reviewed by our science team.  We also tune the base rates to minimize the errors that Ark identified.

I hate to sound like a broken record, but we are trying to reduce the current abuses.  Improvement is not perfection, but improvement is a step in the right direction and better than what we have today. 

[Dan Mangan]

Neal, don't shoot the messenger, but, having spent 20+ years on the marketing/communications side of the technology industry, this is my take...

IMHO, it looks like you sold your soul to Converus.

We're curious... How much did you get for it?

Think it was it a fair bargain?

Think again.

[Neal Harris]

Selling to the USG is no easy task and you have an army of polygraph examiners and related people that will fight you in every possible way. While you may not want to accept it or realize it, Lafayette dropping EyeDetect was not helpful as it gave you instant credibility of sorts in a very specialized market.

Again, I totally agree.  We are well aware of of the aggressive tactics of many USG polygraph examiners.  Thankfully, there are some very good people in government that also want to improve on the current system.  Also, there is congressional pressure on the agencies to justify why their programs are failing so many applicants. 

I already posted that I was disappointed in Lafayette's decision.  I agree that it was not helpful and impacts our credibility.  We should probably issue a press release.  But, I have yet to launch a new technology without a few speed bumps.  It's frustrating, but it is normal for new tech companies to have setbacks.  The hard part is I really like Jennifer and Steve Rider.  I think they are good people that are also frustrated when their products are used improperly.