Olives @olives@qoto.org

While someone doesn't seem to like the point, the point that a few inputs (i.e. people) are unlikely to have a material difference on the outputs of a model remains.

If someone wants to make a copyright or privacy argument, they're free to, but I'm wary of making exaggerations here when that could lead to unintended consequences.

For the record, OpenAI is specifically worse with that, although it is also more restrictive. It's hard to say why. It could be that whatever architecture they use to appear more sophisticated is more prone to this.

In fact, I don't agree with everything that Dr. Tenbergen says, but at least she actually took the time to do a study, and has some years of experience. If she says something interesting, then I might listen.

But, the "AI ethicist" or the "image analysis guy" is not much better than consulting Gary the computer technician about something he hardly knows about.

If someone actually studied it, and they had a bad take, maybe assuming too much bad, then that might be one thing, but you have these people who don't actually know a whole lot (and haven't touched the field) who get consulted as if they have an defining opinion on the matter.

For whatever reason, they keep finding people with hot takes about porn but who don't actually study porn effects, instead they hear something alarming and they decide to echo it.

A guy who has knowledge in analysing images or data is not an expert in sexology.

It might be a step up over an "#AI ethics" person with hot takes about offensive content but their opinion is worth a lot less than someone might make out.

Talking about Sam could be the subject of a long article in itself and I don't have the time or desire to do so.

I don't think you need to be an expert in the human mind to realize that an opt-out for Microsoft Recall is going to lead to a lot of people getting tripped up by it in a sensitive situation (perhaps, one involving security).

If it is going to be there at all, this should be an opt-in instead.

I'm also wondering whether it shouldn't be implemented in some other way, and if there shouldn't be some API or something for a program to opt out of it? #privacy

From what I've heard of it, Microsoft's Recall product is very troublesome from a security perspective.

There is quite a lot for me to read relating to censorship, huh.

Bad caching rears it's ugly head. The sort of caching where old stale content is visible for a long time before being replaced with new content.

I didn't cover Endrass et al specifically (it is referenced indirectly) as it relates to actual child porn, which as we know is not ethical, therefore it's more interesting as part of a suite to debunk anti porn arguments than being interesting in it's own right (where the most it has to say is that someone who views that tend to have different characteristics from actual abusers).

I am aware of it though.

Also, I don't really want to lump people who don't engage in actual child porn in with those who do as I don't think it would be good to invite prejudice upon good people.

I'm reading crazy nonsense conflating teenagers viewing porn with them being "abused" and teens making jokes about sex (or whatever) to each other with them being groomed.

I don't even know how to comment on this. It's too crazy for me. #ukpol

More nonsense for me to read, and maybe comment on.

Palestine. Where someone who might be against hate speech prohibitions suddenly decides they like them when it comes to opposing Israel.

It appears that modifying a video to make it appear that he committed a crime was defamatory.

Cool to see growing attention to the proposed Californian "age verification" bill.

https://qoto.org/@olives/112483833316074543
Out of fear of misguided #DSA / #DigitalServicesAct takes, I refer to my new porn science post (actually a repost of it) where I'm not fond of censorship or "age verification".