|
|
|
|
|
|
by stephen_g
1284 days ago
|
|
|
This paper [1] cites some stats from the UK over one year (I think 2019) - from about 29.4m reports to NCMEC, 102,842 were referred to the NCA, of which 20,038 were referred to policing agencies, which led to 6,500 suspects, and about 750
of those suspects were prosecuted, which it estimates makes up about 3% of all prosecutions for indecent image offences. That comes out to about 0.7% of NCMEC reports to the UK's NCA leading to prosecution. I read another article that linked to stats from Irish and Swiss federal police, and both of them reported discarding about 80% of NCMEC's reports to them in the first stage as not being at all criminally relevant, but I can't find the link right now. 1. https://www.cl.cam.ac.uk/~rja14/Papers/chatcontrol.pdf |
|
|
That doesn’t imply the rest was unactionable. Certainly, the reductions from 20,038 to 6,500 and from there to 750 aren’t due to this tech. The first could be because it produced multiple hits for one person, the second because of policies of the police (minors might have been warned rather than prosecuted, untraceable phone numbers ignored, and persons known to have died removed, for example)
NCMEC also would (rightfully, AFAIK) argue that they never claimed the code to be perfect, and that the earlier reductions are by design.
The real problems, IMO, are (from that pdf):
“the data do not support claims of large-scale growing harm that is initiated online and that is preventable by image scanning.”
and
“The first wave of prosecutions for illegal abuse images, Operation Ore, swept up many innocent men who were simply victims of credit card fraud, their cards having been used to pay for illegal material. A number were wrongly convicted, and at least one innocent man killed himself”