Yes banning ads for misleading health scams would be a good idea, but banning ads for for those who smoke or drink vying for more market share probably isn't purposeful. I'd agree one of the problems of the modern world are young minds are too often introduced far too early to concepts and behaviours they'd cope with better as young adults, but the world has changed and even in everyday life these are in view one way or another.
As for targeting ... I wish I could find the news item from what I think was around 2018-9 that in particular made me sit up and realise the dissatisfaction with wayward algorithms wasn't merely advertising that was not entirely suitable. IIRC this article after the initial introduction of various incidents where the parents had discovered their child had ended up visiting odd places like self harm topics and content not so desirable, the article then outlined someone or a group of people testing various social media areas with new fake accounts. The damming conclusion was just how quickly the algorithms were determining an age and then funnelling apparently somewhat naive youngsters without much prompting into unexpected topic areas, while the fake accounts which suggested an adult experienced something different.
A search today didn't turn up anything close to what was the real concern I believe that got everyone's real attention here in Australia 2018/19, mostly just benign issues like gambling, advertising for alcohol and other marketing opportunities. [1] [2] [4] [5] The govt at the time did ask the platforms to modify what they were doing but the short answer was too hard and not going to do it.
Muddying the waters are long running calls to control platforms by a few interested parties with concerns, some opting to capitalise any time a particular platform faces criticism. Some of these are legitimate long standing problems like [3] where platforms are expected to have live human moderation keeping a watchful eye over their flock at all time ... while ignoring the platforms are generally trying to balance safety with an expectation of privacy. It's these very things that some apparently very powerful parties use to try and bring an and to anonymous content on the web - every time there's an issue trying to install a real ID gateway - many assume the age gateway will do that.
[1] https://www.news.com.au/technology/online/social/leaked-docu... (2017)
[2] https://www.ibtimes.co.uk/facebook-targeting-insecure-childr... (2017)
[3] https://www.forbes.com/sites/zakdoffman/2019/03/01/surge-of-... (2019)
[4] https://www.theguardian.com/technology/2019/oct/09/children-... (2019)
[5] https://insightplus.mja.com.au/2020/48/under-the-radar-preda... (2020)