People around the world are increasingly aware of generative AI and actively using it. According to a survey in six countries—Argentina, Denmark, France, Japan, the United Kingdom, and the United States—the share of people who had heard of at least one of 13 AI tools increased from 78% in 2024 to 90% in 2025. Only 10% of respondents said they had not heard of any. OpenAI's ChatGPT remains the best known, recognized by 73% of people across the countries, followed by Google Gemini at 50% and Meta AI at 46%.
Usage is expanding rapidly. The share of those who have ever used a standalone generative AI system such as ChatGPT rose from 40% to 61%. Weekly usage nearly doubled from 18% to 34%. ChatGPT dominates with 22% weekly users, far ahead of the others. Since May 2024, the user base of each major system has roughly doubled, but most people use them only occasionally—monthly or just once or twice.
Younger people are more active. Among 18- to 24-year-olds, 59% used generative AI in the past week, compared with only 20% of those over 55. This difference is driven mainly by ChatGPT, which has 47% weekly usage among younger people compared with 9% among older people. The differences are smaller for Google Gemini and minimal for Copilot, Meta AI, or Grok because these tools are built into widely used products.
Main Use: Obtaining Information
Obtaining information has become the main reason for using generative AI. Each week, 24% of people use it for this purpose, more than double the 11% in 2024. This has surpassed media creation, which rose by 7 percentage points to 21%. Social interaction, such as using AI as a friend or adviser, is a new category at 7% overall, but 13% among 18- to 24-year-olds compared with 4% among those over 55.
Specifically, tasks such as answering factual questions, which rose from 6% to 11%, or seeking advice are becoming more widespread. The newly added option "researching a topic" stands at 15%. Media creation remains specialized, with growth mainly in image generation, from 5% to 9%, while video (3%) and audio (2%) remained unchanged. Programming or coding remained at 4%, suggesting that early adopters were already doing this in 2024.
Using AI to obtain news doubled from 3% to 6%, but remains a minority activity, particularly in Japan and Argentina. It is more common in Argentina and the United States, among 18- to 24-year-olds (8%) compared with those over 55 (5%), and among people with a university education. Among those who use AI for news, the most common request is for the latest news (54%) or help summarizing, evaluating, or rewriting a story. Younger users are more likely to use it to simplify a story: 48% among 18- to 24-year-olds compared with 27% among those over 55.
Trust in AI and Its Answers in Search
Trust is concentrated among major brands. Across the countries, 29% of people trust ChatGPT, followed by Gemini (18%), Copilot (12%), and Meta AI (12%). Fewer than 10% trust the other brands, mainly because of low awareness. In most countries, more people trust ChatGPT than distrust it, with the United Kingdom being the exception.
AI-generated answers in search are common: 54% saw such an answer in the past week. The highest shares were in Argentina (70%), the United Kingdom (64%), and the United States (61%), and the lowest in France (29%), where Google's AI Overview feature had not yet launched. Among those who saw them, 33% always or often click on source links, 37% sometimes do so, and 28% rarely or never do. Younger people click more often.
Trust in these answers is 50% among those who have seen them, with a neutral minority. Younger adults are more trusting. People value speed, convenience, and the aggregation of information, but trust is conditional in sensitive areas such as health or politics, where many verify information through traditional sources.
Views on AI in Different Sectors and Journalism
People perceive generative AI as ubiquitous: 41% believe it is always or often used across sectors, rising to 67% for search engines and 51% for news media. Views on its impact are mixed but slightly optimistic: 29% expect interactions to improve, while 22% expect them to worsen. Optimists predominate in healthcare, science, and search engines, while pessimists predominate in news, government, and among politicians.
On a personal level, optimists outnumber pessimists in four of the six countries, but when it comes to society, pessimists predominate in three, including the United States. Women expect less improvement for both themselves and society. In journalism, there is a "comfort gap": only 12% are comfortable with fully AI-generated news compared with 62% for entirely human-produced news. With human oversight, this rises to 21%, and with human direction and AI assistance, to 43%.
People are most comfortable with back-end uses such as grammar correction (55%) or translation (53%), and less comfortable with rewriting content (30%) or artificial presenters (19%). More people believe journalists are using AI: the share saying "always or often" increased by at least 3 percentage points across tasks.
Expectations for AI in news are mixed: many anticipate cheaper production (+39 points) and more up-to-date news (+22), but less transparency (-8) and trustworthiness (-19). Views have become more firmly established since 2024. People in Japan and Argentina are more positive, while those in the United Kingdom are more negative. Only 33% believe journalists always or often check AI outputs before publication, with higher figures in Japan (42%) and Argentina (44%) and a lower figure in the United Kingdom (25%). Trust in news is correlated: 57% of those who strongly trust news believe such checks are performed, compared with 19% of those who do not trust news.
43% expect major differences in the responsible use of AI among different news organizations, compared with 28% who expect minor differences. Most people (60%) still do not regularly see AI features such as summaries (19%) or chatbots (16%). Only 19% see AI-generated content labeled daily, even though 77% consume news every day. Only 15% often suspect unlabeled AI use, with the highest share in Argentina (30%).
Use vs. Concerns
According to additional findings from the Thomson Reuters Foundation, AI is transforming journalism worldwide, particularly in the Global South. Journalists there report benefits in speed and efficiency but face challenges involving ethics, equality, and access to technology. Many newsrooms are integrating AI for writing, editing, and translation, but they are concerned about overreliance and bias.
The growth of generative AI is linked to declining trust in institutional journalism and increasing consumption of alternative media through social platforms, complicating the news environment. The Reuters Institute emphasizes three priorities: gathering evidence on AI's impact, supporting the engagement of journalists and editors, and expanding analysis beyond newsrooms to societal consequences, including risks and opportunities.
Source: reutersinstitute.politics.ox.ac.uk
