Trial and error
(In response to Dan Garry’s Sept. 14 column “Unethical detecting.”)
In a recent article, Dan Garry wrote that lie-detecting brain scans should not be admissible as evidence in court if they had "any real possibility of error." This absolutist line of reasoning ignores the fact that the current system has error. A lot.
A statistician from Northwestern University estimated from jury-judge concordance in trial rulings that the error rate in verdicts is at least 12 percent. Even if the current lie-detection technology is only correct 95 percent of the time (and it will almost certainly improve over time), it would still lead to 58 percent fewer incorrect rulings.
In a civilian context, suspects could choose brain-scanning interrogations in lieu of long-lasting, error-ridden, million-dollar trials. Because innocent suspects would benefit from all three of these advantages, they would jump at the chance to be interrogated in this manner.
In practice, the procedure might go something like this: the suspect would, with his lawyer present, be questioned by independent teams using the brain-scanning technology. The blind results from the brain-scanning teams would determine whether the suspect was innocent or guilty. A jury would then sentence the guilty suspects, while taking into account any mitigating circumstances. Additional security could be provided by randomly testing the interrogators and recording all interrogations, fMRI readings, and sentences.
At the prisoner's request, criminal rights watchdog groups could give the data a closer inspection. As this technology improves, all current prisoners could be given the chance to prove their innocence.
This technology has the potential to make our justice system radically more just. Delays in its implementation have the effect of letting the innocent go to jail and the guilty go free.
David Eaton
Reader
