Meta Uses AI Chats to Personalize Ads and Content Across Apps

Meta is beginning to fold users' interactions with its Meta AI conversational features into the data that determines what people see across its family of apps. Starting December 16, 2025, text and voice exchanges with Meta AI, activity from connected surfaces such as Ray Ban Meta smart glasses, and usage from certain AI creative tools will be layered on top of existing signals like likes, follows and post engagements to influence recommendations and targeted advertising across Facebook, Instagram, WhatsApp and Messenger, including Reels and ads.

The company says it will exclude conversations that concern certain sensitive topics from being used as ad targeting signals. Conversations about religious beliefs, political views, health, sexual orientation and racial or ethnic origin are not to be used for personalization, and Meta also says that end to end encrypted WhatsApp and Messenger messages will not be mined for ad signals. The rollout begins in most regions today but will not initially apply in the United Kingdom, the European Union and South Korea.

The change affects a large audience. Meta AI has grown to roughly 1 billion monthly active users across the companys properties, meaning the new signals could rapidly influence recommendations and advertising for a substantial portion of the social networks' user base. Meta will also link AI interactions across apps when accounts are connected, allowing a conversation in one product to alter content and ads in another.

For advertisers the update promises more explicit intent signals. A user who asks Meta AI about hiking boots for beginners could be served ads for footwear, shown local hiking groups, recommended trail Reels or exposed to related posts from friends. The ability to translate conversational queries into audience targeting is widely seen within the industry as a direct path to monetizing generative AI investments.

Privacy advocates and analysts warn the move may blur the line between private assistance and advertising fodder. AI conversations can feel intimate, even when users know they are interacting with a bot, and critics say incorporating those exchanges into ad systems raises new expectations about consent and transparency. Some observers also note differences in how competitors handle chat histories, underscoring the range of approaches to AI and privacy in the industry.

There is no general opt out for users who interact with Meta AI, according to company guidance and third party analyses. Some consumer privacy guides suggest steps such as avoiding Meta AI prompts, submitting opt out requests to the company, subscribing to paid services to reduce personalized ads or deleting accounts, but they caution that these measures may not fully prevent background data collection or the construction of shadow profiles from other information sources.

The change is likely to accelerate the integration of generative AI into digital advertising while intensifying scrutiny from regulators and privacy minded users. As Meta moves conversational data into its recommendation engines, the company will face choices about clarity, limits and oversight if it hopes to maintain user trust while scaling new revenue streams.