Boffoblog: Lying Liars: A Followup

Sun	Mon	Tue	Wed	Thu	Fri	Sat
		1	2	3	4	5
6	7	8	9	10	11	12
13	14	15	16	17	18	19
20	21	22	23	24	25	26
27	28	29	30	31

Lying Liars: A Followup

In my earlier post on the study which reports detecting lying using functional MRIs, I asserted that there is far more media coverage of it than earlier studies, even though it appears to be only an incremental improvement at best over previous ones (and for which there are still open questions about methodology). But I had made that assertion based only on a vague notion that while I remembered reading stories in the past on the subject, I did not know when or how many.

So I decided to do a little Lexis-Nexis search. I will not claim that it was thorough -- the search was of major newspapers over the last two years with the words "brain," "lying" and "MRI" in the full text. Naturally, there were quite a few false positives, but only three stories on point: Financial Times, May 28 of this year; USA Today, November 5 of last year; and Boston Globe, May 1 of last year. None seemed obviously connected to the release of a particular study.

This confirms my hasty assertion, but also raises another topic. The first two stories emphasized the investment made by the CIA and FBI in developing a foolproof lie detection method, an angle mentioned very little in the current spate of articles. I am curious whether such support was behind the new Temple study.

I have no great objection to the role of intelligence agencies developing the technology provided it is accurate; it is also to be hoped that it is in their interest to ensure that such technology is accurate. However, it also ought to be available to non-intelligence agencies and it ought to be publicly vetted.

There is a great leap from saying one can predict at conventional levels of significance whether the subject in an experimental setting is lying, and saying that it is A) of sufficient reliability B) in the real world to drive decisions about witnesses in legal settings, about employees in firms, and anywhere else lie detection might be used. In other words, when questions of civil liberties are relevant, I will not be comfortable if simply at 95 or 99% confidence we can identify differences between liars and nonliars.

Promising technology, but one fraught with potential dangers as well if used prematurely.

04:24 PM in Science | Permalink

The comments to this entry are closed.

Boffoblog

substandard deviations

Recent Posts

Categories

March 2005

Archives

I Recommend:

Bottom of the Page

November 30, 2004

Lying Liars: A Followup

Comments

politics

divertissements

ruminators

sound & vision