Richard Fletcher
@richardfletcher.bsky.social
Director of Research at the Reuters Institute for the Study of Journalism, University of Oxford. Lead researcher + co-author of the Digital News Report. www.richardfletcher.me
Clicking through to links from AI search results is more common among those that trust the responses.
This could suggest that people are using links to find more information rather than check whether AI is accurate - but hard to be sure and probably varies case by case.
This could suggest that people are using links to find more information rather than check whether AI is accurate - but hard to be sure and probably varies case by case.
October 21, 2025 at 7:50 AM
Clicking through to links from AI search results is more common among those that trust the responses.
This could suggest that people are using links to find more information rather than check whether AI is accurate - but hard to be sure and probably varies case by case.
This could suggest that people are using links to find more information rather than check whether AI is accurate - but hard to be sure and probably varies case by case.
Trust in gen AI search responses is relatively high. 50% on average say they typically trust them, compared to 20% who typically distrust.
This is higher than trust in individual chatbots, but mostly because it is more widely used.
Again, younger people have slightly higher trust.
This is higher than trust in individual chatbots, but mostly because it is more widely used.
Again, younger people have slightly higher trust.
October 21, 2025 at 7:50 AM
Trust in gen AI search responses is relatively high. 50% on average say they typically trust them, compared to 20% who typically distrust.
This is higher than trust in individual chatbots, but mostly because it is more widely used.
Again, younger people have slightly higher trust.
This is higher than trust in individual chatbots, but mostly because it is more widely used.
Again, younger people have slightly higher trust.
Only about one third of people say that they always or often click the links to the underlying source that was used to generate the answer - though survey data has limitations here.
As with many behaviours around gen AI, younger people are slightly more likely to say they click through.
As with many behaviours around gen AI, younger people are slightly more likely to say they click through.
October 21, 2025 at 7:50 AM
Only about one third of people say that they always or often click the links to the underlying source that was used to generate the answer - though survey data has limitations here.
As with many behaviours around gen AI, younger people are slightly more likely to say they click through.
As with many behaviours around gen AI, younger people are slightly more likely to say they click through.
There are differences by country, reflecting for example the fact that different platforms in different countries have enabled or disabled AI results for particular types of query (most news queries on Google do not produce AI responses, for example).
October 21, 2025 at 7:50 AM
There are differences by country, reflecting for example the fact that different platforms in different countries have enabled or disabled AI results for particular types of query (most news queries on Google do not produce AI responses, for example).
Following last week's thread on using gen AI for news, this one looks at our research on AI search results based on data from 🇦🇷🇩🇰🇫🇷🇯🇵🇬🇧🇺🇸.
It's already true that most people (54%) say that they have seen AI search results in the last week.
It's more common than the use of all chatbots combined (34%).
It's already true that most people (54%) say that they have seen AI search results in the last week.
It's more common than the use of all chatbots combined (34%).
October 21, 2025 at 7:50 AM
Following last week's thread on using gen AI for news, this one looks at our research on AI search results based on data from 🇦🇷🇩🇰🇫🇷🇯🇵🇬🇧🇺🇸.
It's already true that most people (54%) say that they have seen AI search results in the last week.
It's more common than the use of all chatbots combined (34%).
It's already true that most people (54%) say that they have seen AI search results in the last week.
It's more common than the use of all chatbots combined (34%).
Younger people are using gen AI for news differently.
Older people are more likely to use it for the latest news (red bars) but younger people are more likely to use it to help them navigate news (green).
Gen Z (18-24) are particularly likely to use AI to make a news story easier to understand.
Older people are more likely to use it for the latest news (red bars) but younger people are more likely to use it to help them navigate news (green).
Gen Z (18-24) are particularly likely to use AI to make a news story easier to understand.
October 13, 2025 at 7:35 AM
Younger people are using gen AI for news differently.
Older people are more likely to use it for the latest news (red bars) but younger people are more likely to use it to help them navigate news (green).
Gen Z (18-24) are particularly likely to use AI to make a news story easier to understand.
Older people are more likely to use it for the latest news (red bars) but younger people are more likely to use it to help them navigate news (green).
Gen Z (18-24) are particularly likely to use AI to make a news story easier to understand.
The small group that are using gen AI for news are using it in a variety of different ways - not just as an aggregator.
Though the most widespread use is getting the latest headlines, some are also using it to get more depth, for summarisation, or for making news easier to understand.
Though the most widespread use is getting the latest headlines, some are also using it to get more depth, for summarisation, or for making news easier to understand.
October 13, 2025 at 7:35 AM
The small group that are using gen AI for news are using it in a variety of different ways - not just as an aggregator.
Though the most widespread use is getting the latest headlines, some are also using it to get more depth, for summarisation, or for making news easier to understand.
Though the most widespread use is getting the latest headlines, some are also using it to get more depth, for summarisation, or for making news easier to understand.
Last week I posted a thread on how people are using generative AI - bsky.app/profile/rich...
This thread is on how people use gen AI for news, specifically.
Though getting information is emerging as the key use of gen AI, only 6%on average across 🇦🇷🇩🇰🇫🇷🇯🇵🇬🇧🇺🇸 use it for news weekly, doubling since 2024.
This thread is on how people use gen AI for news, specifically.
Though getting information is emerging as the key use of gen AI, only 6%on average across 🇦🇷🇩🇰🇫🇷🇯🇵🇬🇧🇺🇸 use it for news weekly, doubling since 2024.
October 13, 2025 at 7:35 AM
Last week I posted a thread on how people are using generative AI - bsky.app/profile/rich...
This thread is on how people use gen AI for news, specifically.
Though getting information is emerging as the key use of gen AI, only 6%on average across 🇦🇷🇩🇰🇫🇷🇯🇵🇬🇧🇺🇸 use it for news weekly, doubling since 2024.
This thread is on how people use gen AI for news, specifically.
Though getting information is emerging as the key use of gen AI, only 6%on average across 🇦🇷🇩🇰🇫🇷🇯🇵🇬🇧🇺🇸 use it for news weekly, doubling since 2024.
But if we look at the difference between the proportion that trust and the proportion that distrust we start to see some patterns.
⬆️ A few systems, like ChatGPT, Gemini and Copilot are more trusted than distrusted.
⬇️ Others, like Meta AI and DeepSeek are more distrusted than trusted.
⬆️ A few systems, like ChatGPT, Gemini and Copilot are more trusted than distrusted.
⬇️ Others, like Meta AI and DeepSeek are more distrusted than trusted.
October 8, 2025 at 7:23 AM
But if we look at the difference between the proportion that trust and the proportion that distrust we start to see some patterns.
⬆️ A few systems, like ChatGPT, Gemini and Copilot are more trusted than distrusted.
⬇️ Others, like Meta AI and DeepSeek are more distrusted than trusted.
⬆️ A few systems, like ChatGPT, Gemini and Copilot are more trusted than distrusted.
⬇️ Others, like Meta AI and DeepSeek are more distrusted than trusted.
Finally, some data on trust. It's difficult to interpret data on trust in gen AI systems at the moment because there is limited public awareness - with the exception of ChatGPT, 50% or more have not even heard of them.
October 8, 2025 at 7:23 AM
Finally, some data on trust. It's difficult to interpret data on trust in gen AI systems at the moment because there is limited public awareness - with the exception of ChatGPT, 50% or more have not even heard of them.
In more detail we can see that while most 'getting information' tasks have grown in the last year, most media creation tasks have not - with the exception of using generate AI to create images.
October 8, 2025 at 7:23 AM
In more detail we can see that while most 'getting information' tasks have grown in the last year, most media creation tasks have not - with the exception of using generate AI to create images.
What are people using gen AI for?
We've seen a significant shift in the last year, with using AI to get information (e.g. answering factual questions) overtaking using it to create media (e.g. writing emails, making images).
Will this be the main use of gen AI in the future?
We've seen a significant shift in the last year, with using AI to get information (e.g. answering factual questions) overtaking using it to create media (e.g. writing emails, making images).
Will this be the main use of gen AI in the future?
October 8, 2025 at 7:23 AM
What are people using gen AI for?
We've seen a significant shift in the last year, with using AI to get information (e.g. answering factual questions) overtaking using it to create media (e.g. writing emails, making images).
Will this be the main use of gen AI in the future?
We've seen a significant shift in the last year, with using AI to get information (e.g. answering factual questions) overtaking using it to create media (e.g. writing emails, making images).
Will this be the main use of gen AI in the future?
No surprises that there's an age gap in the use of gen AI, with use more widespread among younger people.
But this only applies to standalone systems like ChatGPT. For AI embedded in other products, like Meta AI and Copilot, there's no age gap because the host product is used by people of all ages.
But this only applies to standalone systems like ChatGPT. For AI embedded in other products, like Meta AI and Copilot, there's no age gap because the host product is used by people of all ages.
October 8, 2025 at 7:23 AM
No surprises that there's an age gap in the use of gen AI, with use more widespread among younger people.
But this only applies to standalone systems like ChatGPT. For AI embedded in other products, like Meta AI and Copilot, there's no age gap because the host product is used by people of all ages.
But this only applies to standalone systems like ChatGPT. For AI embedded in other products, like Meta AI and Copilot, there's no age gap because the host product is used by people of all ages.
On the use of specific systems, ChatGPT is still the most widely used (22% weekly) - ahead of Gemini (11%), Meta AI (9%) and Copilot (6%).
Worth remembering that the public's use of a lot of the tools favoured by experts, like Claude and Perplexity, is still very marginal - 1% weekly.
Worth remembering that the public's use of a lot of the tools favoured by experts, like Claude and Perplexity, is still very marginal - 1% weekly.
October 8, 2025 at 7:23 AM
On the use of specific systems, ChatGPT is still the most widely used (22% weekly) - ahead of Gemini (11%), Meta AI (9%) and Copilot (6%).
Worth remembering that the public's use of a lot of the tools favoured by experts, like Claude and Perplexity, is still very marginal - 1% weekly.
Worth remembering that the public's use of a lot of the tools favoured by experts, like Claude and Perplexity, is still very marginal - 1% weekly.
A thread on how people's use of generative AI has changed in the last year - based on survey data from 6 countries (🇬🇧🇺🇸🇫🇷🇩🇰🇯🇵🇦🇷 ).
First, gen AI use has grown rapidly.
Most people have tried out gen AI at least once (61%), and 34% now use it on a weekly basis - roughly doubling from 18% a year ago.
First, gen AI use has grown rapidly.
Most people have tried out gen AI at least once (61%), and 34% now use it on a weekly basis - roughly doubling from 18% a year ago.
October 8, 2025 at 7:23 AM
A thread on how people's use of generative AI has changed in the last year - based on survey data from 6 countries (🇬🇧🇺🇸🇫🇷🇩🇰🇯🇵🇦🇷 ).
First, gen AI use has grown rapidly.
Most people have tried out gen AI at least once (61%), and 34% now use it on a weekly basis - roughly doubling from 18% a year ago.
First, gen AI use has grown rapidly.
Most people have tried out gen AI at least once (61%), and 34% now use it on a weekly basis - roughly doubling from 18% a year ago.
Those on the left and right are typically less likely than those in the centre to think their political views are covered fairly - but the results vary a lot by country.
In the US those on the right are much less likely to think their views are covered fairly, but this is not a common pattern.
In the US those on the right are much less likely to think their views are covered fairly, but this is not a common pattern.
August 11, 2025 at 11:33 AM
Those on the left and right are typically less likely than those in the centre to think their political views are covered fairly - but the results vary a lot by country.
In the US those on the right are much less likely to think their views are covered fairly, but this is not a common pattern.
In the US those on the right are much less likely to think their views are covered fairly, but this is not a common pattern.
In most countries women are less likely than men to think that coverage of people their gender is fair.
August 11, 2025 at 11:33 AM
In most countries women are less likely than men to think that coverage of people their gender is fair.
Perceived fairness of coverage of "people my age" tends to have a curved relationship with age - so the youngest and oldest are less likely to think coverage is fair than those in middle age.
August 11, 2025 at 11:33 AM
Perceived fairness of coverage of "people my age" tends to have a curved relationship with age - so the youngest and oldest are less likely to think coverage is fair than those in middle age.
New paper with @kirstenaeddy.bsky.social in Journalism on whether people across 45 countries think news coverage of people their age, their gender, and with their political views is fair or unfair.
We outline some interesting demographic differences (more details in 🧵)...
doi.org/10.1177/1464...
We outline some interesting demographic differences (more details in 🧵)...
doi.org/10.1177/1464...
August 11, 2025 at 11:33 AM
New paper with @kirstenaeddy.bsky.social in Journalism on whether people across 45 countries think news coverage of people their age, their gender, and with their political views is fair or unfair.
We outline some interesting demographic differences (more details in 🧵)...
doi.org/10.1177/1464...
We outline some interesting demographic differences (more details in 🧵)...
doi.org/10.1177/1464...
Final @reutersinstitute.bsky.social presentation at #ICA25 today.
@amyrossarguedas.bsky.social with Digital News Report data on who listens to podcasts, as part of a panel on amplifying communication research through podcasts.
@amyrossarguedas.bsky.social with Digital News Report data on who listens to podcasts, as part of a panel on amplifying communication research through podcasts.
June 15, 2025 at 2:12 PM
Final @reutersinstitute.bsky.social presentation at #ICA25 today.
@amyrossarguedas.bsky.social with Digital News Report data on who listens to podcasts, as part of a panel on amplifying communication research through podcasts.
@amyrossarguedas.bsky.social with Digital News Report data on who listens to podcasts, as part of a panel on amplifying communication research through podcasts.
3. @felixsimon.bsky.social and Ralph Schroeder in a panel on what currently know (and don't know) about AI in journalism. #ICA25
June 14, 2025 at 1:38 PM
3. @felixsimon.bsky.social and Ralph Schroeder in a panel on what currently know (and don't know) about AI in journalism. #ICA25
2. @academiuhhh.bsky.social @atrimukherjee.bsky.social me and @antoniskalog.bsky.social on the effect of social media vs TV in reducing news use gaps. #ICA25
June 14, 2025 at 1:34 PM
2. @academiuhhh.bsky.social @atrimukherjee.bsky.social me and @antoniskalog.bsky.social on the effect of social media vs TV in reducing news use gaps. #ICA25
Three @reutersinstitute.bsky.social presentations at #ICA25 today.
1. @waqasejaz.bsky.social Mary Sandford and me on the link between news use, climate anxiety and pro-environment behaviour across 8 countries.
1. @waqasejaz.bsky.social Mary Sandford and me on the link between news use, climate anxiety and pro-environment behaviour across 8 countries.
June 14, 2025 at 1:34 PM
Three @reutersinstitute.bsky.social presentations at #ICA25 today.
1. @waqasejaz.bsky.social Mary Sandford and me on the link between news use, climate anxiety and pro-environment behaviour across 8 countries.
1. @waqasejaz.bsky.social Mary Sandford and me on the link between news use, climate anxiety and pro-environment behaviour across 8 countries.
2. Tali Aharoni, me & @rasmuskleis.bsky.social with a poster on a five wave panel study in 🇬🇧 on the link between news use and trust in news in the first weeks of the COVID pandemic. #ICA25
June 13, 2025 at 2:04 PM
2. Tali Aharoni, me & @rasmuskleis.bsky.social with a poster on a five wave panel study in 🇬🇧 on the link between news use and trust in news in the first weeks of the COVID pandemic. #ICA25
Two #ICA25 presentations from @reutersinstitute.bsky.social researchers today.
1. For the early risers, me @felixsimon.bsky.social @waqasejaz.bsky.social & @rasmuskleis.bsky.social on what variables are associated with positive/negative attitudes towards the use of generative AI in news in 🇬🇧🇺🇸🇩🇰🇦🇷🇯🇵🇫🇷
1. For the early risers, me @felixsimon.bsky.social @waqasejaz.bsky.social & @rasmuskleis.bsky.social on what variables are associated with positive/negative attitudes towards the use of generative AI in news in 🇬🇧🇺🇸🇩🇰🇦🇷🇯🇵🇫🇷
June 13, 2025 at 2:04 PM
Two #ICA25 presentations from @reutersinstitute.bsky.social researchers today.
1. For the early risers, me @felixsimon.bsky.social @waqasejaz.bsky.social & @rasmuskleis.bsky.social on what variables are associated with positive/negative attitudes towards the use of generative AI in news in 🇬🇧🇺🇸🇩🇰🇦🇷🇯🇵🇫🇷
1. For the early risers, me @felixsimon.bsky.social @waqasejaz.bsky.social & @rasmuskleis.bsky.social on what variables are associated with positive/negative attitudes towards the use of generative AI in news in 🇬🇧🇺🇸🇩🇰🇦🇷🇯🇵🇫🇷