On 7/26/23 19:45, John Levine wrote:
I've heard that working on CSAM is really difficult, even traumatic, work. So I can understand why they might have tunnel vision. But people with tunnel vision should not be dictating policy.Warning about the dangers of _introducing_ end-to-end encryption to messaging apps, as Apple threatens to withdraw already-end-to-end secure Messages and FaceTime from the UK, is very much relying on and appealing to a non-technical, non-critical thinking, audience.I know people who work on CSAM and while they are uniformly working hard to fight it, they do tend to a degree of tunnel vision and unfortunate assumptions that anyone who makes it harder for them to do their jobs is ignorant or malicious. And then there's the "nerd harder and give us a back door only good people can use" stuff.
I am skeptical of such conclusions in the absence of a thorough and sober, balance-of-harm analysis. For instance, it should not be automatically assumed that back doors or other CSAM countermeasures will actually reduce the level of CSAM that's produced or distributed. (People often naively assume that new solutions will work ideally, that they won't be circumvented, etc.). And of course surveillance can impose tremendous costs on both individuals and society, especially when that surveillance is misused or the controls established for use of those surveillance tools are altered or compromised (which experience says is extremely likely at least in the long term).But that can cut both ways. There is absolutely a lot of bad stuff that is passed through encrypted channels, and shrugging and saying too bad, can't do anything is not going to make us any friends. I agree that on balance the benefits of encryption outweigh the costs, but the costs are real.
So, basically, without extensive, comprehensive, detailed, and publicly-reviewed analysis, an argument that any particular technological measure will improve the CSAM situation is unpersuasive.
Keith
