A man spent a year in jail on murder charge that hinged on disputed AI evidence
1–10 of 211 posts
Re: A man spent a year in jail on murder charge that hinged on disputed AI evidence
#2This is about ShotSpotter.
Re: A man spent a year in jail on murder charge that hinged on disputed AI evidence
#3The article says that employees of the AI company (ShotSpotter) manually reviewed and classified the sounds as gunshots:
> records showed that ShotSpotter actually initially picked up what sounded like a firework a mile away, and this was later reclassified by ShotSpotter staff to be a gunshot at the intersection where and when Williams was seen on camera.
So the AI didn’t even make the call. The staff did, manually. I assume that means the actual audio is available and entered into evidence?
If humans are making the call then blaming AI seems like a stretch. That’s almost like blaming the motion detection algorithms for triggering video recordings that were later reviewed by humans. It’s still humans reviewing the recordings and making decisions.
Re: A man spent a year in jail on murder charge that hinged on disputed AI evidence
#4The article says that employees of the AI company (ShotSpotter) manually reviewed and classified the sounds as gunshots: > records showed that ShotSpotter actually initially picked up what sounded like a firework a mile away, and this was later reclassified by ShotSpotter staff to be a gunshot at the intersection where and when Williams was seen on camera. So the AI didn’t even make the call. The staff did, manually.…
It may not be right to blame AI, but it's certainly the case that they were presenting it as "the system says there was a shot fired here." The bigger point is that the "system" can't be trusted. And the fact that they need to manually classify things is proof of that fact. (It's way more complicated than just an audio recording. It's recordings from numerous locations, attempting to triangulate the original location, without being confused by echoes and other things.)
Re: A man spent a year in jail on murder charge that hinged on disputed AI evidence
#5Well that's still a failure on both the software programmed by humans causing a false positive and then doubled down by a wrong human review, based on a false alert.
All involved stakeholders have failed.
Re: A man spent a year in jail on murder charge that hinged on disputed AI evidence
#6The article says that employees of the AI company (ShotSpotter) manually reviewed and classified the sounds as gunshots: > records showed that ShotSpotter actually initially picked up what sounded like a firework a mile away, and this was later reclassified by ShotSpotter staff to be a gunshot at the intersection where and when Williams was seen on camera. So the AI didn’t even make the call. The staff did, manually.…
That's the point? These systems invariably turn out to not be any sort of AI at all, but sales people ready to make a call any way their customer (police) wants it made.
Later it's presented as AI to a jury. It's basically the same scam as lots of forensics, junk science to get a desired result.
Re: A man spent a year in jail on murder charge that hinged on disputed AI evidence
#7The article says that employees of the AI company (ShotSpotter) manually reviewed and classified the sounds as gunshots: > records showed that ShotSpotter actually initially picked up what sounded like a firework a mile away, and this was later reclassified by ShotSpotter staff to be a gunshot at the intersection where and when Williams was seen on camera. So the AI didn’t even make the call. The staff did, manually.…
It may not be right to blame AI, but it's certainly the case that they were presenting it as "the system says there was a shot fired here." The bigger point is that the "system" can't be trusted. And the fact that they need to manually classify things is proof of that fact. (It's way more complicated than just an audio recording. It's recordings from numerous locations, attempting to triangulate the original location…
> It may not be right to blame AI, but it's certainly the case that they were presenting it as "the system says there was a shot fired here."
I think a crucial distinction is to clarify who “they” is: Who was making the claim? From the article, it seems that the prosecutors may have been the ones exaggerating the evidence:
> Prosecutors said ShotSpotter picked up a gunshot sound where Williams was seen on surveillance camera footage in his car, putting it all forward as proof that Williams shot Herring right there and then.
Naturally the defense wanted to see the actual evidence, so that’s what they got down to.
I know everyone wants to vilify AI, but I think the actual bad actor in this article might be the prosecutor trying to misrepresent the evidence.
Re: A man spent a year in jail on murder charge that hinged on disputed AI evidence
#8It's interesting to me that there's no working process to introduce/accept new types of science in court. Polygraphs aren't allowed (though the police still find ways to use it to bully people). So, there's some process. But you still see stories about subsets of fiber, hair, arson, fingerprint, dna, etc, science, that shouldn't have been brought to court.
Re: A man spent a year in jail on murder charge that hinged on disputed AI evidence
#9Coming next year: "A man spent a year in jail on child pornography charge that hinged on disputed Apple CSAM evidence".
Re: A man spent a year in jail on murder charge that hinged on disputed AI evidence
#10Well that's still a failure on both the software programmed by humans causing a false positive and then doubled down by a wrong human review, based on a false alert. All involved stakeholders have failed.
How is this a failure by software? The software categorized the sound as a fire work.