«Hype does not match interest»: only 1%-7% of people use ChatGPT daily, up to 30% do not know about it — research

Large corporations are now investing heavily in generative artificial intelligence and models such as ChatGPT. A new study Reuters Institute and Oxford University shows that very few people actually use artificial intelligence, despite the tremendous hype around this technology.

An online survey of 12,000 people in six countries (Argentina, Denmark, France, Japan, the UK, and the US) focused on whether people use generative artificial intelligence and what they think about its application in journalism and other fields. Below are some of the results.

• ChatGPT is the most recognized generative AI, with about 50% of respondents having heard of it. It is also the most widely used generative AI tool in six countries. However, only 1% use ChatGPT on a daily basis in Japan, 2% — in France and the UK, and 7% in the US. Many of those who claim to have used it have done so only once or twice.
• While there is a general awareness of generative AI, between 20% and 30% of people have not heard of any of the most popular AI tools. Younger people are more likely to use generative AI on a regular basis. On average, across all six countries, 56% of people aged 18-24 have used ChatGPT at least once, compared to 16% of people aged 55 and older. 5% use generative AI to get the latest news.
• A significant portion of the public believes that journalists are currently using generative AI for certain tasks. 43% believe that they always or often use it to edit spelling and grammar, 29% — to write headlines, and 27% — to write the text of an article.
• When asked whether news created by mostly human-supervised artificial intelligence should be labeled, the vast majority of respondents want at least some disclosure or labeling. Only 5% of our respondents say that none of these use cases should be disclosed. About one-third think that editing the spelling and grammar of an article (32%) and creating a headline (35%) should be labeled, and about half think that writing the article (47%) and analyzing the data (47%) should be labeled.
• Most people expect generative artificial intelligence to have a major impact on virtually all areas of life over the next five years. From 51% expecting a significant impact on politics, to 66% expecting a significant impact on media, and 66% expecting a significant impact on science.

Overall, most people expect generative AI to improve their lives. When asked whether generative AI will make society as a whole better or worse, people were generally more pessimistic. Concerns about artificial intelligence are mainly centered on privacy and security.

Dr. Richard Fletcher, lead author of the report, said BBCThere is a mismatch between the hype around AI and public interest in it: «A significant part of the population is not particularly interested in generative artificial intelligence».