Researchers say parent company Meta is failing to remove explicit images on the social media site
Summary
A Danish study found Instagram’s moderation of self-harm content to be “extremely inadequate,” with Meta failing to remove any of 85 harmful posts shared in a private network created for the experiment.
Despite claiming to proactively remove 99% of such content, the platform’s algorithms were shown to promote self-harm networks by connecting users.
Critics, including psychologists and researchers, accuse Meta of prioritizing engagement over safety, with vulnerable teens at risk of severe harm.
The findings suggest potential non-compliance with the EU’s Digital Services Act.
To be clear the last ban was a ByteDance ban, not a TicTok ban. If ByteDance sells off it's portion of TikTok, TikTok would be able to remain.
But in this case, the Section 230 needs to re-visited. It was written before social networks started algorithmically deciding what to show people. It needs to be adapted so these companies can be held liable for the content they "recommend" to users. Once they can be sued for their algorithm's decisions, those algorithms will magically get better. Or at least less harmful.
And if they can be sued for algorithmic recommendations at all, I feel like their resort would be going back to pure "user" recommendation like community managed feeds, following other's recs and such. Which would kill SO MANY birds with one stone.
How AI comes into the picture is that Meta claims that they do this via "AI", which cannot be known how it works, so they can't be blamed for it being ineffective, and the research group put together a piece of software that did a better job in 3 days, so Meta should either git gud or stop bullshitting about trying their best.
Centralized, engagement-driven, advertisement financed, social media, is the internet's greatest mistake.
Prior to that shit, the internet was pretty great. The fediverse is way to fix much of it.