2 ms·
If it's an actual ML model then someone can leak it and develop countermeasures. If it's a contractor who claims to be AI but is actually some guy in a room in
by astrange 4y ago
If it's an actual ML model then someone can leak it and develop countermeasures. If it's a contractor who claims to be AI but is actually some guy in a room in Maryland, that's not possible but also the results wouldn't be very reliable.
Of course, the article just says they're planning to do this. It doesn't say it's going to work, and our closest examples, forensic science like handwriting and blood spatters and polygraphs and all that, generally don't actually work.