- Weekly generative AI use is growing rapidly: Half of respondents (51%) across six countries surveyed consistently now use standalone generative AI systems weekly, up from 34% in 2025.
- ChatGPT remains the market leader: Weekly use increased to 34%, while Google Gemini grew to 24%, Microsoft Copilot to 11% and Claude to 7%.
- AI is increasingly used to find information: The proportion using generative AI for information-seeking rose from 11% in 2024 to 38% in 2026 across six countries.
- Public acceptance of AI-generated news remains low: Only 12% of respondents across nine countries are comfortable with news produced entirely by AI, increasing to 19% when human oversight is involved.
- Transparency and trust are major newsroom challenges: Respondents expect AI to make news cheaper and more up to date, but also less transparent and less trustworthy.
Generative AI is becoming an increasingly routine part of everyday life, with half of respondents in six countries now using standalone AI systems weekly. However, public enthusiasm for the technology has yet to translate into greater acceptance of its use in journalism.
These are among the findings of the Reuters Institute for the Study of Journalism’s 2026 report on generative AI and its impact on society, based on a YouGov survey across nine countries: Argentina, Australia, Denmark, Finland, France, Japan, Norway, the UK and the US.
Rapid growth
The research points to rapid growth in both intentional AI use and exposure to AI features embedded in existing digital platforms. Across the six countries surveyed consistently since 2024, weekly use of standalone generative AI systems increased from 34% in 2025 to 51% in 2026. Daily use has nearly doubled year-on-year, rising from 16% to 28%.
ChatGPT remains the market leader, with weekly use increasing from 22% to 34% across the six countries. However, competitors are gaining ground. Google Gemini rose from 11% to 24%, Microsoft Copilot from 6% to 11%, and Anthropic’s Claude from 1% to 7%.
AI exposure extends well beyond people actively choosing to use chatbots. Some 57% of respondents across the nine countries encountered AI-generated answers, summaries or assistants in search engines such as Google and Bing during the previous week. A further 25% encountered AI on their devices and 21% through social media or messaging apps.
News organisations are also contributing to this exposure. Almost half (49%) had encountered at least one AI feature on a news website or app in the previous week. AI-generated content labels were the most commonly noticed feature (24%), followed by bullet-point news summaries (20%).
AI becomes an information intermediary
One of the report’s significant findings is the growing use of AI to find information. Across the six countries with comparable historical data, the proportion using generative AI to obtain information in the previous week increased from 11% in 2024 to 38% in 2026.
Researching a topic remains the most common application (24%), while seeking advice (20%) is now almost as prevalent as asking factual questions (18%).
Using AI specifically for news remains relatively uncommon, although it is increasing. The proportion doing so rose from 3% in 2024 to 9% in 2026 across the six countries, with the figure at 8% across all nine.
Among those using AI for news, asking follow-up questions about a story (49%) has overtaken getting the latest news (46%) as the most common activity.
Convenience is a major driver. Almost two-thirds (64%) say they use AI because it offers faster, easier answers, while 56% value the interactive nature of conversational interfaces. Only 21% say they use these systems because they provide more reliable information than other sources.
Trust remains a challenge for journalism
Despite growing familiarity with AI, attitudes towards its use in newsrooms remain broadly negative and have changed little since 2024.
Across the nine countries, just 12% say they are comfortable with news being produced entirely by AI. This increases to 19% when there is some human oversight.
Respondents are generally more comfortable with AI being used for supporting tasks such as translation and correcting spelling and grammar than for producing news content. They are particularly uncomfortable with AI-generated commentary and opinion.
The perceived trade-offs are equally revealing. Across six countries, respondents expect AI to make news cheaper to produce (+34 percentage points) and more up to date (+19 points), but less transparent (-10 points) and less trustworthy (-17 points).
The report also identifies a wider divide in public expectations about AI. People are generally more optimistic about its use by scientists and healthcare professionals, but more pessimistic about its impact when deployed by governments, social media companies, politicians and news organisations.
Importantly, the researchers find that scepticism is driven more by concerns about honesty and transparency around AI use than by perceptions of competence or good intentions.
The findings present a challenge for news publishers: while audiences are increasingly encountering and using AI in their everyday information-seeking, they remain cautious about allowing it to take a more prominent role in journalism.
For an industry exploring AI-driven efficiencies, products and new ways of reaching audiences, public confidence and transparency remain essential considerations.
The story is based on the supplied executive summary and introduction. The research covers nine countries and should not be presented as representative of South African public opinion.
Read the report here.
