> Don't get me wrong: I am against ChatControl. Because of one argument I believe to be valid: we fundamentally cannot know what the algorithm doing the scanning is doing, so those who control it could abuse it. Of all the discussions I have seen against ChatControl, I haven't seen another valid argument. But this one is enough.

It is not enough to know what the algorithm is doing. It also needs to be possible (for the average user as well) to stop it from doing reprehensible things. If a client-side scanning algorithm is actually searching for e. g. political content, it is possible to detect it via reverse engineering, but merely knowing it won't solve the problem, but instead lead into self-censorship.