(The latter still account for the ads I see most of the time, unfortunately.)
What is being targeted is surveillance-based advertising methods. These involve the collection, brokering and combining of user data. This data is purchasable by anyone - including US government agencies which have been using it as way of obtaining information without oversight(1). There is an expectation that other governments and bad actors are also obtaining this data for advantage.
This type of advertising is also responsible for poorly targeted ads that follow you around the internet. Perhaps you mentioned something in passing on an instagram chat, or you liked a photo from a friend on holiday.
Consumers generally underestimate their digital footprint and the risks associated with having this information available. It's more information than what we'd trust our own governments possessing in a single, or any, database, yet we let others take it without any oversight whatsoever. Additionally the information gathered about them can be wrong or invade their privacy in ways they aren't expecting (E.g. infer their sexuality or private desires) (2). Furthermore individual users can be targeted which beyond being able to prank someone(3), is also ripe for exploitation. (4)
(1) https://www.documentcloud.org/documents/23844477-odni-declas... or the easier to read: https://www.nbcnews.com/tech/security/us-government-buys-dat...
(2) https://techcrunch.com/2018/05/16/facebook-faces-fresh-criti...
(3) https://www.adweek.com/performance-marketing/roommate-makes-...
(4) https://techcrunch.com/2021/10/15/researchers-show-facebooks...