Absolutely this. It’s not about banning content - to a certain extend parents are responsible for what their kid has access to. In many cases proper education allows kids to self regulate and consume “adult” content appropriately with no harm. Unfortunately that’s not possible everywhere as many countries lack the resources and/or mentality to achieve this.
Regardless, dangerous content and services (just like dangerous substances) should be hard to make and very visibly marked, leaving no doubt about what it is and how it works. I love the EU’s notion of “algorithmic transparency” [0]. I would go a step further for systems attempting to increase engagement by exploiting behavioural sensitivities to be marked and even opt-in (think cigarette packaging).
To sell drugs, cigarettes, alcohol etc, you need somebody to do the selling, and that person needs to be located wherever your customers are. If you break that place's laws, well, they probably have police who can put you in jail.
Social media is different, you can run a social media website targeted at Americans without ever stepping foot in the US, having a server in the US, having a business entity in the US etc. It's just some random Russian website, following Russian but not American laws, that some Americans like to visit. Sure, the US can try playing the cat-and-mouse game with you and force ISPs to block your site and all VPNs, proxies and TOR nodes that might presumably give customers access to it, but that's still a game one needs to play, and playing it isn't without consequences for the privacy and freedom of others, consequences most democratic governments aren't willing to bear.
Even if the government wins and somehow manages to block you completely, or put enough obstacles in your path that doing further business doesn't make sense, you still don't lose. You don't go to jail or face legal action, you just quit and focus on other countries instead.
Regardless, dangerous content and services (just like dangerous substances) should be hard to make and very visibly marked, leaving no doubt about what it is and how it works. I love the EU’s notion of “algorithmic transparency” [0]. I would go a step further for systems attempting to increase engagement by exploiting behavioural sensitivities to be marked and even opt-in (think cigarette packaging).
[0] https://algorithmic-transparency.ec.europa.eu/index_en