Recently OpenAI released some research that tracked how people globally have been adopting and using ChatGPT from its launch (November 2022) through to July 2025.

It uses a representative sample of over 1 million actual ChatGPT conversations (all classified via a privacy-preserving automated pipeline) and examines who is adopting ChatGPT (demographics, occupation, country) and how usage has shifted over time.  

Here’s a summary of what we know about how it relates to news.

By July 2025, ChatGPT had been adopted by around 10% of the world’s adult population. Growth has been especially fast in lower-income countries. Early adopters skewed male, though that gender gap has narrowed substantially over time. A large portion of messages come from users under 26. 

Perhaps unsurprisingly, those with more education, in knowledge-intensive or highly paid professions are more likely to use ChatGPT for work tasks. This is important for news organisations to know because there is also a correlation between education and propensity to pay for news. 

Interestingly, non-work related messages are growing faster than work-related ones. By mid-2025, more than 70% of all messages are non-work related. Work usage is more common among users with higher education and those in higher-paid professional jobs. 

So what do people use ChaGPT for? This is where it gets interesting. Take a look at the chart below:

Three topics dominate: practical guidance, seeking information, writing. Together, they make up nearly 80% of conversations. 

News specifically isn’t broken down, but it likely falls into the 18.3% of requests asking for “specific info.” This means it may comprise a decent chunk of user questions. On the upside, this is great for news as it gives more leverage. On the flipside, we don’t yet have strategies/business models for consumers moving to answer engines.    

Health, fitness, beauty, or self-care at 5.7% of queries may provide news organisations with some relevant, valuable content.

Also note that while we might expect commerce to grow, it’s pretty small at 2.1%. My hunch is that better commerce products will be launched which may tick this up, but we have no information around that. 

Clearly we don’t know how this compares to other answer engines, but it’s a useful glimpse into the consumer needs. 

What does that mean for us?

1. We are competing for daily habit, not just attention. If 70%+ of ChatGPT use is now non-work related, people are building everyday habits with AI — asking about health, cooking, personal guidance, entertainment.
2. Answer engines eat into information discovery “seeking information,” which directly competes with search and Q&A functions that once drove news traffic.
3. Writing and content creation is a core value. News organisations can use AI, as many workers do for writing, editing, summarising, and translating. They could also use it to provide AI-assisted services such as summaries, translations, or contextual explainers, building on their trust and authority.
4. Opportunity for younger markets. Adoption growth is strongest among under-26 users, demographics where many news organisations struggle with monetisation.
5. Shifting competitive benchmark. The competitive question for publishers is: What problem am I solving for the user, and do I do it better than AI? That means leaning into trust, authority, community, and original reporting — areas AI cannot easily replicate.
6. Data and personalisation expectation: Answer engines raise expectations. Audiences will expect answers tuned to their context (location, interests, habits). News publishers must invest in first-party data and contextual delivery or risk irrelevance.

However you look at this, ChatGPT is shifting user behaviour from “searching” to “asking” and from “consuming” to “doing.” News media must decide whether to integrate with this new behaviour, compete through differentiation, or both.

If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .