Meta announced in early October 2025 that starting December 16, they’ll use your conversations with their AI to target advertisements on Facebook and Instagram. This includes data collected through Meta’s AI chat features across WhatsApp, Messenger, Instagram, and notably, their Ray-Ban Meta smart glasses. If you’re outside the EU, UK, or South Korea, there’s no way to opt out short of stopping use of Meta’s AI features entirely.
The move marks a significant shift in how social media platforms monetize AI technology. While other platforms like ChatGPT and Google’s Gemini use conversation data only to improve their models, Meta is directly connecting your private chats to its advertising engine.
What Data Meta Will Collect
Meta’s new policy covers any interaction you have with their AI assistant across their family of apps. This means casual questions, planning conversations, or brainstorming sessions can all influence what ads you see.
The Ray-Ban Meta smart glasses add another dimension to this data collection. These glasses include a 3K camera and microphone that record whenever you say “Hey Meta.” By default, they store these voice recordings for up to a year. The camera captures photos and 3-minute videos hands-free. If you use cloud-based AI features to analyze scenes or images, those pictures and videos get uploaded to Meta’s servers and may be used for both AI training and personalized ads.
Meta has removed the ability to disable voice recording by default on these glasses. Recordings now happen automatically unless you manually delete them in settings.
What’s Off Limits (and What Isn’t)
Meta says they’ll exclude a narrow set of sensitive topics from advertising. Conversations about religion, political views, health, sexual orientation, ethnicity, race, philosophical beliefs, or trade union membership won’t be used to target ads.
Everything else is fair game. Leisure interests and lifestyle preferences can all be used for ad targeting. If you chat about hiking plans, Meta’s systems will infer you like hiking and start showing you ads for hiking gear and related products. Vacation ideas, sports interests, shopping preferences, fitness goals, entertainment tastes, and family events could all influence your ad profile.
The Privacy-Convenience Trade-Off
This approach creates a clear trade-off. On one hand, AI-driven targeting makes ads more relevant. When done well, you see fewer random ads and more that actually match your interests. If you’re researching camping equipment in a private chat, seeing ads for tents and hiking boots could feel helpful rather than intrusive.
On the other hand, mining private conversations feels invasive. The algorithm might misinterpret context, or data could be exposed if leaks occur. As privacy experts point out, AI systems are extremely data-hungry and users have even less control over what gets collected.
The approach amounts to pervasive surveillance of daily life through your private conversations. While AI targeting may improve ad relevance, it subjects intimate personal signals to Meta’s advertising engine.
How This Differs from Google and OpenAI
Other AI platforms handle conversation data very differently. OpenAI’s ChatGPT explicitly states that chat content is used only to train and improve models, not to profile users for advertising. Their policy is clear: “We do not use your content to market our services or create advertising profiles.” ChatGPT users can even opt out of model training through privacy controls.
Google’s Gemini operates similarly. It retains chat transcripts briefly, uses the data to improve the AI, but explicitly states “Gemini Apps chats are not being used to show you ads.”
Meta is the outlier here. While Google and OpenAI treat conversation logs as training data (with opt-outs available), Meta treats those same logs as fresh signals for its advertising business.
Cross-Platform Data Sharing
A key risk in Meta’s ecosystem is that data flows across all their apps when accounts are linked. Meta’s “Accounts Center” connects your Facebook, Instagram, and WhatsApp profiles. If you’ve linked these accounts, a conversation in one app can influence ads in another.
Meta confirms this in their privacy policy: “We use information across Meta Company Products from the accounts that you choose to add to the same Accounts Center.” A WhatsApp AI chat about travel plans could show up as targeted travel ads on your Facebook feed if both accounts are linked.
If you keep accounts separate (not in the Accounts Center), Meta says it won’t share that AI chat data between platforms. The privacy risk grows with account linking.
Can You Disable This?
Meta has provided no direct on/off switch for this new ad targeting feature. Outside protected regions, users are opted in by default. Both Meta and tech reporters confirm “there is no way to opt out” of having your AI chat data used for ads.
The company points only to existing ad-preference tools as user controls, but none of those settings stop the collection of AI conversation signals. In practice, the only way to prevent it is to stop using Meta’s AI features or their apps entirely.
European and UK users are exempt due to stricter GDPR rules on consent and profiling. For everyone else, unless Meta adds a new privacy option, there’s no way to disable AI-based ad personalization without abandoning Meta’s services.
What This Means for Users
We’re heading toward a world where the line between our personal lives and the algorithms that track us keeps disappearing. If devices and platforms already know everything about us, they’ll probably be more helpful. The algorithms will understand context better and provide more relevant suggestions.
But there’s a cost. That cost is privacy. Your private thoughts, shared in what feels like a personal conversation with an AI assistant, become data points in an advertising profile.
Most people will probably accept this trade-off, especially as the technology improves and the convenience becomes more apparent. But some will push back, seeking alternatives or limiting their use of these features.
The Regulatory Divide
Meta’s policy doesn’t apply in the EU, UK, or South Korea due to stricter data protection laws. Under GDPR, using personal conversation data to target ads typically requires explicit consent, which Meta hasn’t implemented in those regions.
UK authorities recently confirmed that users have a right to object to targeted advertising. In a recent case, the UK’s Information Commissioner’s Office backed a citizen’s demand that Meta stop profiling them for ads, making clear that people must have a clear way to refuse such data use.
In the US, no single law explicitly bans this kind of profiling yet. Some state laws, like California’s CCPA, allow consumers to opt out of certain data sales, which could arguably cover AI chat data if classified as sensitive. However, compared to the EU and UK, US regulation is less prescriptive here.
Meta’s policy proceeds in the US and other regions without an explicit opt-out requirement for now. Future rulemaking or lawsuits mirroring the UK case could force changes if regulators decide personalized AI ads violate consumer privacy expectations.
What Business Owners Should Know
If you’re a business owner advertising on Meta’s platforms, this change could improve your ad targeting significantly. Your ads may reach people who’ve expressed genuine interest in related topics through their AI conversations.
However, it’s worth considering the bigger picture. As AI algorithms get better at understanding context and intent, the advantage of traditional advertising diminishes. People will increasingly rely on AI assistants to make recommendations and decisions for them.
The businesses that win in this new landscape won’t just be the ones with the best ads. They’ll be the ones that AI algorithms consistently recommend because they genuinely provide the best products, services, and customer experiences.
If you want your business to show up when people ask ChatGPT, Google’s Gemini, or Meta’s AI for recommendations, the work starts now. You need to make sure your brand information is comprehensive, accessible, and authoritative across the web.