AI error jails innocent grandmother for months in North Dakota fraud case
(www.grandforksherald.com)
What kind of fucking court would accept AI output as evidence, even for probable cause? Where has it ever been validated? What are the documented Type 1 and Type 2 error rates in independent peer-reviewed tests? Where is the output that gives the rationale for why it said there was a match? Because without that, there's no more reason to believe it than a Magic 8-Ball.
And holy shit, she's 50? That's three miles of bad Tennesee road.