Meta’s AI is summarizing some bizarre Facebook comment sections

If you’ve been on the Facebook app recently, you may have seen Meta’s AI inject itself into the comments section with summaries of what people say. Given how wild Facebook comment sections often get, it’s not hard to imagine how ridiculous some of these roundups become. (By the way, this isn’t the first time Meta’s AI has appeared in the comments section: 404 Media I saw him pretending to be a father in a Facebook group.)

After seeing screenshots of the feature shared on Topics It is Reddit, I decided to check the comments sections on my Facebook app. I found AI summaries appearing in many of the posts I checked – confusing responses and all. An AI summary on a post about a store closure said: “Some commenters attribute the closure to the store being ‘woke’ or poor selection, while others point to the increase in online shopping.”

Another Facebook post from Deputy about Mexican street fighters sparked a comment section summary that said some people were “less impressed” with the performance and referred to it as a “stupid way to ask for a handout.” The AI ​​also captured some of the lightest jokes people made about a bobcat sighting in a Florida town. “Some marveled at the sighting, with one commenter hoping the lynx remembered the sunscreen.”

It’s not yet clear how Meta chooses which posts to display comment summaries on, and the company did not immediately respond. On the edgerequest for comment.

Anyway, the summaries don’t really include anything that I would find useful (unless you like vague notions about what random people have to say) – but it might help you identify posts where the comments section has become too toxic to bother scrolling.

AI summaries have also raised privacy concerns, as Meta is feeding user comments into its AI system to generate them. Over the past week, many Facebook and Instagram users in the European Union and the United Kingdom have received a notification that Meta will train its AI on their content. (Data protection laws in both regions require Meta to disclose this information.) While Meta allows these users to object to the use of their data to train AI, the process is not that simpleand the company has rejected some users requests.

Here in the US, Meta’s privacy policy page says the company uses “shared information about Meta’s products and services” to train AI, including posts, photos and captions. Meta allows you to submit a request to correct or delete personal information used to train its AI models, but this only applies to third-party information. Everything else seems to be fair game.