Artificial intelligence will read your conversations! This will apply to Messenger, WhatsApp, and Instagram

  • Starting December 16, 2025, Meta will begin using the content of conversations with Meta AI for ad personalization
  • The change will affect over a billion people monthly – roughly one in eight people on the planet
  • Opt-out is not possible; exceptions apply only to sensitive topics such as health or religion

Sdílejte:
Jakub Kárník
Jakub Kárník
3. 10. 2025 02:30
meta-facebook-soukromí
Advertisement

Meta has announced a significant change to its privacy policy, which will affect all users of Facebook, Instagram, and WhatsApp. Starting December 16, 2025, the company will analyze the content of conversations with Meta AI and use it to personalize ads and content recommendations. Simply put: if you chat with the chatbot about hiking, you will start seeing ads for hiking boots.

What exactly is changing

Meta claims that the change is intended to „improve content recommendations so that you see what truly interests you, and less of what doesn’t.” In practice, this means that every text and voice conversation with Meta AI will be analyzed and used to create your advertising profile.

An example Meta itself provides: if you chat with AI about hiking, you might subsequently start receiving recommendations for hiking groups on Facebook, friends’ posts about hikes, or ads for hiking footwear. This is the same principle Meta has been using for years for your posts, likes, or followed pages – only now, conversations with artificial intelligence will be added to the mix.

Cannot opt out, only adjust consequences

Meta explicitly states that you cannot opt out of this change. The only way to prevent the analysis of your conversations is not to use Meta AI at all. However, if you are already using the chatbot, you do not have the option to prohibit the use of this data for ads.

While you can still adjust which ads you see through the Ads Preferences settings, this only mitigates the consequences – it won’t stop the conversation analysis itself. Paradoxically, Meta offers control over the outcome but not over the process itself.

Exceptions for sensitive topics

Meta promises that conversations about sensitive topics will not be used for displaying ads. Protected topics include:

  • Religious beliefs
  • Sexual orientation
  • Political opinions
  • Health status
  • Racial or ethnic origin
  • Philosophical beliefs
  • Trade union membership

WhatsApp has an exception – but only a partial one

Meta states that if you do not have a WhatsApp account linked via Accounts Center to Facebook or Instagram, conversations with Meta AI on WhatsApp will not be used for ad personalization on other platforms. This is good news for users who keep WhatsApp separate.

However, this does not mean that Meta does not analyze WhatsApp conversations at all – it just won’t use them across platforms. Meta still has access to the content of these conversations and can use them in other ways, for example, to train its AI models.

A billion users as guinea pigs

Meta states that Meta AI is used by over a billion people monthly. Considering that there are approximately 8 billion people on the planet, this means the change will affect roughly one in eight people on Earth. This is not a marginal group of users – it is a massive real-time privacy experiment.

Users will be notified of the change starting October 7, 2025 via in-app notifications and email. There are several weeks between the notification and the actual implementation of the change, but Meta offers no way to decline the change – it merely informs you that it will happen.

What this means in practice

Meta presents this change as an „improvement to the user experience,” but the reality is simpler: it gains another data source for ad targeting. While until now Meta analyzed your posts, likes, followed pages, or searches, it will now also add what you chat about with the chatbot.

For example, if you search for information about cars, Meta has already recorded that. But if you chat with Meta AI about considering buying an electric car and having a budget of up to $30,000, Meta suddenly knows much more – not only that you are interested in cars, but also what type, in what price range, and at what stage of decision-making you are. This is a goldmine for advertisers.

While Google and OpenAI also train their AI models on user data, Meta goes a step further – not only does it use your data to improve AI, but it directly links it to your advertising profile.

It is also important to note that while Meta AI has arrived in Europe this year, it speaks Czech quite poorly and sometimes deliberately avoids our language. Therefore, in the Czech Republic, probably a negligible percentage of users actually use Zuckerberg’s chatbot, but this will likely change in the future with increasing development.

Do you use Meta AI?

Source: Meta, GSMArena

About the author

Jakub Kárník

Jakub is known for his endless curiosity and passion for the latest technologies. His love for mobile phones started with an iPhone 3G, but nowadays… More about the author

Jakub Kárník
Sdílejte: