On Thu, 30 Sep 1999, Matthew Gaylor wrote:
> Am I the only one to wonder about the use of increasingly abstruse
> "scientific" evidence in criminal trials?
> Example: DNA samples. What, precisely, prevents prosecutors from
> simply printing out graphics and columns of figures that support
> their contention? If they are trying to "prove" that DNA samples
> taken from a crime scene match sequences from their suspect, what's
> to prevent them simply showing two copies of the same slide, side by
> side? Could you tell the difference? To accomplish it wouldn't even
> entail much of a "conspiracy:" one person along the chain of evidence
> could effect the switch, and it would be hard to trace in the face of
> the automatic/reflex government stonewall.
There was a presentation by the director of the FBI crime lab at the TIGR genome meeting discussing the use of DNA forensics in criminal cases. She discussed how in the early days, there was a lot of mishandling but now the FBI has issued handling standards and is working with local authorities to train people properly. DNA forensics has also been reviewed and "certified" by the National Academy of Sciences. My basic feeling from the dicussion is that the techniques are fairly robust at this point.
Nothing in the world is going to keep evidence from being planted occasionally. But this is good cop/bad cop and one would hope that we have more of the former than the later.
> Evidence such as DNA matching is functionally "sourceless," since
> even if you were provided complete documentation on it, you still
> couldn't understand it: unless you're already an expert, you have no
> common-sense way to judge whether it's valid or not.
That isn't true. If it is documented from multiple witnesses as coming from the crime scene and the paper trail is good (as it should be with any evidence), the lab tests now are of the quality and the math has been extensively analyzed such that this is a legitimate form of evidence. I would trust a DNA test far more than I would trust a ballistics report. Both can be forged but both can be easily reexamined (so forging them doesn't do much good).
Now what was of interest is the database that the FBI is curating involving DNA fingerprints from known criminals. Apparently the rationale is that when you become a criminal, you give up the "privacy" of your appearence or fingerprints. Since your DNA is also an identifying feature, it goes into the database as well. Now this is good from the perspective that it is allowing connections to be made in criminal cases that would not otherwise be made. She cited a case of a rapist/murderer who was also a drug courier who showed up as committing crimes in multiple states based on the DNA matches). She cited something like 11 cases where people who had been convicted of a crime had been released because DNA tests indicated they were not the real criminal involved.
However, there are significan privacy concerns with regard to the DNA database. Currently if the access to the database abused by local governments (it would be interesting to find out what such abuse would be), the FBI is supposed to revoke the access.
It was of interest that she did not know the answer to the question of whether such a database could or would be used for genotyping for "violence" genes. And she certainly didn't have an answer for the question I had regarding what would happen if such genes were discovered in the database -- would such individuals be considered "disabled"?
Its going to get very interesting in the future when "special interest groups" start getting concrete knowledge with regard to their nature imposed "disabilities"...
To fan the flame wars... What if the right to carry a gun required you to have your genome "edited" replacing those violence genes with non-violent variants?
Would a better solution to lengthy imprisonments be engineering criminals for passivity or pacifistity?