A real truth machine?

Mon, 14 Dec 1998 09:20:30 -0600

Thanks to Scott Badger for his comments and his suggestion that I give you some background on my own feasibility studies of a Truth Machine. I apologize for taking so long to respond. Writing novels is a sideline/hobby with me, my real career being numismatics (dealing in rare coin). Along with manufacturers and sellers of electric generators, freeze-dried foods, and, I suspect, weapons and ammunition, rare coin dealers have found themselves the sudden beneficiaries of an expanding clientele dropped into our laps by the current and escalating Y2K preparedness craze. My company has been inordinately busy, and as a result some of my other interests have suffered.

At the time I was researching The Truth Machine (1993 through 1995), I came to the conclusion that polygraphs were only about 60 to 80 percent accurate, depending to a great extent on the skill of the operator. I believed then, as I do now, that computer systems similar to those that now approve or reject most American Express credit card transactions (far more successfully than the best human operators ever could) will be able to heighten that percentage significantly. And accuracy could be tweaked up even further, I believe, if combined with voice-stress analysis and perhaps other corollaries to deceit. But even 99% accuracy would fall far short when contrasted against the benefits of 100% deceit detection depicted in my novel.

Since The Truth Machine was first published, I've received dozens of letters from people claiming to have invented perfect lie detectors. Most were based on ridiculous concepts. One method that particularly resonates involved recording statements on a tape-recorder, then playing them backwards and somehow analyzing the vibrations. But there was one fellow with very impressive credentials who claimed to have conducted 100% successful field tests using a very similar technology to that described in my novel (MRI and/or PET scans, I think, run through a $25,000 NeXT computer.) He was looking for funding, and I even considered making an investment, but decided against it because, before demonstrating the machine, he wanted to me to sign a confidentiality agreement that read suspiciously like a precursor to a lawsuit. I remain skeptical that such a machine is possible with only today's technology, yet this man was by far the most credible person who contacted me claiming success. I've scanned my files for his address, but as yet have not located it. I do remember that the man's name was Dr. Michael Tansey, and I'm fairly sure he was from New Jersey. I will conduct a more thorough search when time permits, and if I find any more information, I'll post it here in case anyone is interested in getting in touch with him.

Unfortunately, I continue to believe, as my novel portrayed, that the construction of a 100% accurate deceit-detector is an astonishingly complex project.

The notion of a "deceit center" in the brain is still theoretical, yet this seems to be the most likely approach to building a foolproof lie detector. Professor Daniel Schacter, Chairman of the Psychology Department at Harvard, published a brain scanning study a few years ago that showed differences in brain activity during true and false recognition. (Professor Schacter also wrote an excellent book called Searching for Memory, which delves into many of the nuances that differentiate false memory from actual deceit. I recommend it highly. Available at amazon.com.) Unfortunately, as Professor Schacter recently explained to me, this test had nothing to do with deceit. Rather it was designed to detect differences in brain patterns when recalling accurate memories versus false memories. Potentially useful, but not nearly as valuable, in my opinion, as detecting intentional deceit.

A brief note to Greg Burch: I agree that the "ubiquitous surveillance of human activity" discussed by Brin in his book, The Transparent Society (also highly recommended), has the potential to reduce human mischief. But I maintain that a foolproof lie detector would be a necessary adjunct to such surveillance. Without it, in a world of unimaginably powerful future technologies, a person who gains trust through such surveilled actions might find him/herself in a position to end a substantial subset - or perhaps all - of human life.

It's chilling to think about, but I haven't changed my mind: A universally available method to detect malicious intentions seems humanity's only sure safeguard against such potentially destructive technologies as will inevitably exist very soon.

Best regards,
Jim Halperin

<http://www.heritagecoin.com> <http://www.firstimmortal.com>
<http://www.truthmachine.com> Or to purchase or post a review of The
First Immortal or The Truth Machine:
<amazon.com/exec/obidos/ASIN/0345421825/jameslhalpe> or