Conversations are the new clicks

On 16 December 2025, Meta will begin using data from users’ AI chat interactions to refine its ad targeting. The change applies to almost everywhere except the EU, UK, and South Korea, where privacy laws still block such use.

It’s a defining moment in the evolution of digital advertising: conversations with AI assistants are not private chats with something you think you control, but data streams feeding the world’s biggest marketing machine.

For years, Meta has relied on clicks, likes, and views to understand its users. Now it’s adding something far more revealing, the context of your conversations. Chatting with Meta AI about holiday plans or hobbies will soon translate into targeted ads for flights, hiking boots or recipes. Meta describes this as improving ad relevance. But what it really signals is a shift in how first-party data is defined. Conversations are the new clicks.

In our Horizons session, back in 2024, we talked about how everyone will have an AI assistant in the near future and now we are here keeping our data under our control has encountered another challenge.

Being a media agency, we are on the side of more data should equal more value. Delivering more relevant ads in suitable environments, to help pay for the service you are using, should be a win for everyone. 

But we know keeping our data ours and seeing intrusive and/or awful ads will always be a thing. 

With the direction of less control of your data and AI creating really poor ads that inexplicably play the happy birthday tune, the fear advertising is going to become even more invasive and even more annoying, is very real.

The line between helpful and invasive has never been so paper thin. Meta claims sensitive topics like religion and health are excluded, but for most other subjects, the game is on.

There’s no way to opt out unless you avoid using the AI tools altogether. The supposed trade-off is convenience for surveillance.

Even Meta’s paid subscription, which removes ads, doesn’t guarantee your privacy. You may not see ads, but your data still fuels the engine. It’s the illusion of control, you pay not to be targeted directly, but your information remains monetised indirectly. Privacy, in the age of AI assistants, is extremely unlikely.

Maybe we need Blockchain to intervene…?  Absolute control over your own data, where you can allow access, licence it even and revoke cleanly as you choose.

Obviously, this development raises profound ethical questions. AI assistants are tools to profile us more than ever before.  How you vote, eat, work and play have been influenced by your digital habits for years, but now you’re openly chatting into the abyss, giving it all of the context and it doesn’t need to infer the context from a click, it can respond back to make sure.

Chat feels personal, almost confidential, but every exchange is another line in your behavioural profile. That sense of intimacy is what makes this next phase of data collection so powerful.

Away from Meta, OpenAI is already exploring monetisation for ChatGPT beyond subscriptions, including advertising and sponsored content. Google’s Gemini, naturally integrated into search and YouTube, has all the ingredients to use conversational data for ad targeting. Microsoft and Amazon will be watching closely, each with ecosystems that would benefit from similar data insights.

Every AI assistant wants to be your daily companion. But as these systems become central to how we learn, shop, and make decisions, they also become prime territory for behavioural targeting. When your chat about a new hobby leads to a perfectly timed product suggestion, the convenience will feel seamless. But behind that seamlessness lies the same incentive that’s driven the web for two decades: turning human attention into data and data into revenue.

You used to think Facebook listened to your conversations via your phone, well now you’re actively talking directly to Facebook, via your phone.

But let’s face it, this is just how good advertising works.  It can use this data to drive value for you, or it can use this data to influence how you think.  The data is now more powerful, as it’s based on a 2-way conversation, rather than inferred from a click.

The trusted PHD in your pocket is now grassing you up for cash, so someone can sell you some trainers.

Further reading:

Meta: Improving Your Recommendations on Our Apps With AI at Meta

Forbes: Meta AI To Start Using Your Data For Targeted Ads—Here’s How And When

Tech Crunch: Meta plans to sell targeted ads based on data in your AI chats