The UK's "Online Harms" bill will grievously harm privacy and
freedom
of speech.
Article 19 also warned of a
threat to freedom of speech.
There is currently a moral panic in the UK about a teenager who
committed suicide, apparently under the influence of
antisocial
media sites' recommendation algorithm.
When they saw she was looking at postings of despair, they directed
her to more despair, which became overwhelming.
This suggests another approach which could achieve the goal without
threatening people's freedom of speech or people's privacy: namely, to
restrict the recommendation algorithms only.
The most basic way is to make a rule that the site must not offer any
recommendations, any "promoted" posts, except when the user makes an
explicit request — "please show me something you think might interest
me". Even if the site's algorithm for deciding what to show you when
you request this is unchanged, it will operate less often and show
you fewer things.
The next step might be to have a way for users to mark a posting as
"not nice". A posting thus marked would still be on line, so you
could still see it if you ask for it specifically, but the site would
never recommend it automatically to anyone.
We can conceive of many variations of this idea. Surely some variation
of them will avoid the real danger without causing real repression.