Meta CEO Mark Zuckerberg holds a smartphone as he makes a keynote speech at the Meta Connect annual event at the company’s headquarters in Menlo Park, California, on Sept. 25, 2024.
Manuel Orbegozo | Reuters
Meta will soon show ads and other content to users based on their interactions with the company’s digital assistant and related products powered by generative artificial intelligence.
The social media giant announced the update to its recommendation system on Wednesday, and said it will go into effect on Dec. 16. Users will receive notifications of the change starting on Oct. 7.
The move underscores how Meta is attempting to better tie its billion-dollar investments into generative AI with its core online advertising business.
Meta spent the summer on a major AI hiring and spending blitz, and said in July during its second-quarter earnings report that the AI initiatives will “result in a 2026 year-over-year expense growth rate that is above the 2025 expense growth.”
The company currently provides generative AI abilities to users via its Meta AI digital assistant, which is bundled across its apps like Facebook, Instagram, WhatsApp and Messenger. Meta AI is also available as a stand-alone app and website.
Users can interact with Meta AI like they do with OpenAI’s ChatGPT, asking it to provide directions or generate images based on prompts.
The Mark Zuckerberg-helmed company also said that the Meta AI digital assistant now has more than 1 billion active monthly users, although that figure is not specific to the Meta AI stand-alone app.
Zuckerberg said in May that Meta AI reached 1 billion monthly active users, and added that eventually “there will be opportunities to either insert paid recommendations” or offer “a subscription service so that people can pay to use more compute.”
Meta privacy and data policy manager Christy Harris said during a media briefing that people already assumed that the Facebook parent was incorporating generative AI interactions with ad targeting and content recommendation.
“While this is a natural progression of our personalization efforts and will help give us even better recommendations for people, we want to be super transparent about it and provide a heads up before we actually begin using this data in a new way, even if people already thought that we were doing this,” Harris said.
Harris provided an example of how the update could influence the kinds of content and ads people may see across Facebook, Instagram and other Meta-related apps.
When users chat with Meta AI about how to plan their family vacations, the digital assistant’s responses could influence the kinds of Reels short videos they may be recommended to watch on Facebook, for instance.
“So the Reels that I see on my Facebook feed or other types of content that is recommended to me could include family friendly travel destinations,” Harris said. “It could include ads for hotels or other signals that would be informed by the conversation that I have had with Meta AI.”
People’s voice interactions with the Meta AI digital assistant when wearing the company’s Ray-Ban Meta glasses will also influence the company’s recommendation engine, Harris said.
“So whether you’re using your keyboard to type your interactions or you are using the audio version of an interaction, those signals will still be used,” she said.
The recommendation update will not take into account user interactions with Meta AI via WhatsApp unless people link their accounts on the messaging service to other sibling apps like Instagram, Harris said.
People will not be able to opt out of the upcoming recommendation engine change, but if they don’t interact with Meta AI, the “update won’t apply to them,” Harris said.
Meta plans to debut the recommendation update to users in the U.K. and the EU “following our usual regulatory updates,” Harris said.
WATCH: How to save the Internet.
