> “We cannot live with incorrect information sending people to prison.”
Don't trust data given by third parties then.
Don't trust data given by third parties then.
It appears data came from the telecom companies and straight to the police's IT department. I'm not sure what parties you're referring to?
Unless you're suggesting not to use data from telecom companies? In which case how would they get any records at all?
Yes, I'm suggesting to not use data from telecom companies, unless they have a way to prove that said data is real.
Maybe in this specific case the telecom data was correct, but can you assure the same for every datum ever provided?
Trusting telecom data is essentially the same as trusting an eyewitness.
The eyewitness says the truth sometimes, but you can't trust them as an infallible source of truth.
There's some evidence that this is the case. In Germany we had a very high profile case where the police used contaminated swabs for taking DNA samples. There's also evidence that both DNA and fingerprint attribution is not trustworthy (from a strictly scientific perspective).
This gets imho even worse with data. Even in good faith an ISP can't be trusted. There is no chain of custody for the location records, they are very likely faulty, not immutable and so forth. As an accused, I have no way of challenging the records, unlike other forensic evidence like DNA tests.
Not only contaminated, the swaps weren't even rated for lab use. They spend years looking for a phantom serial killer active in half the country because several police labs were cutting costs. The mysterious killer was working in a swab factory.
> There's also evidence that both DNA and fingerprint attribution is not trustworthy (from a strictly scientific perspective).
Isn't the issue that they use so few markers that you would be guaranteed collisions the moment you tried to build a large scale biometric database?