Author: admin

This report is based on a survey conducted between August and November 2024 with a broadly representative sample of 1,004 UK journalists. The survey was primarily focused on whether and how journalists and news organisations use artificial intelligence (AI), and how it relates to other aspects of their work. Our analysis of the data produced the following key findings:

On whether and how UK journalists and their newsrooms use AI:

• More than half (56%) of UK journalists use AI professionally at least once a week, another 27% use it less frequently, and only 16% have never used it for journalistic tasks.
• UK journalists most frequently use AI for language-processing tasks, specifically ‘transcription/captioning’ (49% of UK journalists use AI for this task at least monthly), ‘translation’ (33%), and ‘grammar checking/copy-editing’ (30%).
• AI is also used for more substantive journalistic tasks such as ‘story research’ (22% use AI for this task at least monthly), ‘idea generation/brainstorming’ (16%), ‘generating parts of text articles (e.g. headlines)’ (16%), ‘fact-checking/verification/source assessment’ (12%), and ‘generating first drafts of text articles’ (10%).
• ‘Generating parts of text articles (e.g. headlines)’ and ‘story research’ come equal fourth out of 31 journalistic tasks when those tasks are ranked according to the proportions of UK journalists using AI to perform them on a daily basis.
• Few UK journalists use AI for audio or video generation – only 4% and 2% respectively do so at least monthly.
• In terms of broad task groups, journalists in the UK use AI most frequently for ‘initial newsgathering’ followed by ‘information processing or analysis’, ‘journalistic writing or rewriting’, ‘backend production or management tasks’, and ‘audio or image/graphic generation or editing’.

On which UK journalists use AI:

• Younger journalists and journalists identifying as male use AI professionally somewhat more frequently.
• Journalists with higher levels of management responsibility use AI professionally more frequently.
• Professional AI use is linked to a journalist’s reporting beat. For example, 43% of business journalists use AI professionally at least weekly compared with 21% of lifestyle journalists.
• After controlling for age and gender, we found no statistically significant associations between journalists’ contract type (e.g. permanent, fixed term, freelance) and the frequency with which they worked with AI in their journalistic tasks, indicative of how access to some AI tools has been democratised.
• Being involved in the production of journalism in any of three media formats – ‘text’, ‘graphics, cartoons, illustrations, or animation’, and ‘video’ – was associated with more frequent AI use. Involvement in the production of photographs reversed that association.
• The more media formats UK journalists produced in, the more frequent their professional use of AI.

On links to job satisfaction:

• UK journalists who are more frequent AI users are more likely to believe they work on low-level tasks too frequently.
• UK journalists who are more frequent AI users are not more satisfied with the amount of time they work on complex and creative tasks.

On UK journalists’ attitudes towards AI:

• UK journalists tend to be pessimistic about the effect of AI on journalism: 62% see it as a ‘large’ or ‘very large’ threat to journalism, and only 15% see it as a ‘large’ or ‘very large’ opportunity.
• Almost all groups in the data are more likely to see AI as a threat than as an opportunity but more senior journalists, those with higher levels of AI knowledge, and those that regularly use AI are more likely than average to see it as an opportunity (and usually less as a threat).
• More than half of UK journalists are ‘extremely concerned’ about the potential negative impact of AI on public trust in journalism (60%), on the value of accuracy (57%), and on the originality of journalistic content (54%). They are less concerned about the inadvertent exposure of personal data (25%).
• Differences in levels of concern between different demographic groups are small (5 percentage points or less), but concern is higher among those with more AI knowledge and considerably lower among those that use AI for journalistic tasks on a daily basis.

On levels of AI integration in newsrooms:

• Most UK journalists (60%) say that there has been some AI integration in their newsroom, but integration is overwhelmingly described as ‘limited’, with very few reporting extensive or full integration.
• Differences between news outlet types are small, but AI integration is more extensive in newspapers, commercial media (as opposed to publicly owned media), and conglomerates (as opposed to independent outlets).
• UK journalists overwhelmingly expect their main outlet’s use of AI to increase in the future (63 percentage point difference between those that think it will increase vs decrease), and are more likely to describe their outlet’s stance on AI to be supportive (39%) rather than opposed (20%).

On news outlets’ approaches towards AI:

• Most UK journalists (60%) say their main news outlet has established AI protocols or guidelines around at least one of the issues we asked about, such as ‘human oversight and control’ (44%), ‘data privacy and security’ (43%), and ‘transparency’ (42%).
• UK journalists are less likely to say their main news outlet has established protocols or guidelines around AI ‘bias and fairness’ (27%).
• Around one third of UK journalists (32%) say that their main news outlet provides training on AI technologies.
• Most UK journalists (57%) say that their main news outlet only uses third-party AI tools, with fewer reporting that they only use tools developed in-house (9%) or a combination of both (34%).
• UK journalists whose main outlet is publicly owned or part of a conglomerate –disproportionately broadcasters or newspapers – are more likely to say their outlet has established AI protocols and guidelines, provides AI training, and uses AI tools developed in-house.

1. Whether and how UK journalists and their newsrooms use AI

Many prior studies on whether and how journalists use AI have explored the topic qualitatively, interviewing relatively small samples of news workers. Although such studies provide valuable insights, they do not attempt to generalise about AI use among journalists more broadly – to do so requires larger and more representative samples. Some larger surveys of journalists’ AI use have now taken place. However, their foci on generative AI (GenAI) rather than AI more widely, their use of unrepresentative samples, or how they did not distinguish between current and future use means we still have a limited understanding of whether and how journalists are using AI professionally. By surveying a broadly representative sample of (UK) journalists about their actual use of AI – including but not limited to GenAI – we hope this study will help fill this knowledge gap.

This first section reveals how frequently UK journalists are using AI and for which specific tasks. It shows that the application of AI is most common in language-processing tasks (e.g. transcription) but is also happening with more substantive journalistic tasks, such as story research. Although we find a majority of UK journalists use AI weekly in a professional context, use varies by gender, age, reporting beat, level of management responsibility, and the media formats – such as text and photographs – journalists work with. The section ends with evidence that AI has not yet fulfilled the hopes some had that it would relieve journalists from low-level tasks and enable them to feel fully satisfied with the variety, challenge, and creativity of their work.

1.1 Overall frequency of AI use by UK journalists

In this report we adopt a broad definition of AI, one that encompasses a range of technologies from rule-based automation to more advanced machine-learning-based systems. In order that our respondents understood what we consider AI in journalism to include, our definition was prominently displayed in the survey. It read:

We define AI technology as the automation of tasks or decisions that would previously have required human intelligence, such as identifying trends, producing content, and personalising news, using techniques like machine learning, natural language generation, and rule-based automation

(Thurman and Thäsler-Kordonouri 2025).

With this definition fresh in their minds, we asked journalists about the specific ways in which they use AI professionally, showing them a list of 31 tasks and asking them how often they work with AI when performing each task.

The results show that 27% of UK journalists make use of AI for at least one journalistic task on a daily basis, 29% do so 1–4 times a week, 16% 1–3 times a month, and 11% less than once a month. The other 16% said they had never worked with AI on a journalistic task (see Figure 1).

Figure 1.

Prior to this question we had asked the general question ‘How frequently do you personally work with AI in your journalistic tasks?’ The answers to this question show lower levels of use. ^{This is likely due to how prompting respondents with a list of specific tasks that can be AI-supported increased recall.}

1.2 The specific uses UK journalists are making of AI

Figure 2 shows how often, on average, UK journalists use AI for 31 specific journalistic tasks. The three most frequent uses are for what we might call language processing, specifically transcription/captioning (49% use AI for this task at least monthly), translation (33%), and grammar checking/copy-editing (30%).

One explanation for why these language-processing tasks top the list may be that the accuracy problems associated with AI output (see, for example, Kalai et al. 2025) are of less concern in these contexts than where AI is used for tasks, such as fact-checking and reporting breaking news, that are more at ‘the core of journalistic labour’ (see, for example, Cools and Diakopoulos 2024: 14).

Nevertheless, our findings clearly show that UK journalists are also using AI for substantive journalistic tasks. For example, 22% of journalists use AI at least monthly for story research; 16% for idea generation/brainstorming; 16% for generating parts of text articles (e.g. headlines); 12% for fact-checking/verification/source assessment; and 10% for generating first drafts of text articles (see Figure 2).

Figure 2. 

Indeed, when we order the 31 tasks by the proportions of journalists who use AI on a daily basis when performing them, generating parts of text articles (e.g. headlines) and story research come equal fourth (4% use AI daily) after grammar checking/copy-editing (14%), transcription/captioning (8%), and translation (5%).

At the other end of the scale, our results show that AI is rarely used by UK journalists for still image or video generation. Only 6% and 2%, respectively, of journalists use AI for these tasks at least monthly (see Figure 2). Some news organisations have specifically ruled out using AI in this way, such as the New York Times, which promised to ‘not use artificial intelligence to manipulate photos or videos or use the technology to generate images to represent real events or situations’ (NYT Trust Team, 2024).

Such a stance aligns with audience expectations. A study by the Reuters Institute for the Study of Journalism, using representative population samples from Argentina, Denmark, France, Japan, the UK, and the US, found that acceptance of AI-generated audio-visual news content is low: 74% of respondents were not in favour of news organisations ‘creating a realistic image when no photo exists’, and an overwhelming 81% did not support the use of ‘AI-generated presenters/authors’ (Simon et al. 2025: 8).

As with video generation, our findings show that AI use for comment moderation and page layout is also relatively rare. This is likely because these are specialist tasks not carried out by many journalists.

To provide a high-level perspective, we grouped the 31 individual tasks into five broad categories:

• initial newsgathering
• information processing or analysis
• journalistic writing or rewriting
• audio or image/graphic generation or editing
• backend production or management.

When comparing AI usage across these categories, our findings show that initial newsgathering tasks are most frequently supported by AI, followed by information processing or analysis tasks, and journalistic writing or rewriting tasks. Apart from transcription/captioning, the tasks in the backend production or management category are infrequently AI-assisted. Taken together, the tasks in the audio or image/graphic generation or editing category are least supported by AI (see Figure 3).

Figure 3.

1.3 Specific uses of AI at UK journalists’ main news outlet

Our survey not only asked journalists about their own professional use of AI, but also about use in their/their employer’s main news outlet. To measure the specific ways AI is being used in UK journalists’ workplaces, we showed our respondents a list of tasks. This list contained all the tasks we had used when we questioned journalists about their own professional use, plus three additional tasks – revenue generation (e.g. dynamic paywalls, personalised subscription pitches); making personalised content recommendations to users; and customer relationship management (e.g. chatbots) – that are usually carried out centrally by a small number of specialists. Our respondents were asked whether or not, excluding their own use, the main news outlet they worked for used AI for each task.

Unsurprisingly, tasks journalists commonly see AI being used for in their main news outlet are tasks they report using AI for frequently too. For instance, language-processing tasks top the news outlets list, with 34% of journalists stating that AI is used for transcription/captioning at their main news outlet. Translation (17%) and grammar checking/copy-editing (16%) come second and third (see Figure 4).

Interestingly, although story research was the fourth most commonly AI-assisted task when we asked journalists about their own professional practice (see Figure 2), the task was eighth in the list of tasks journalists had seen AI being used for at their main news outlet. One explanation could be that some journalists do not feel comfortable sharing their AI use on this substantive task with their colleagues.

Figure 4.

1.4 Which UK journalists use AI?

1.4.1 Age and gender

Our survey finds that the use of AI by UK journalists is associated with some of their demographic characteristics, employment conditions, and professional practices. ^{Both age and gender play a role in how frequently UK journalists use AI professionally. Specifically, younger journalists and those who identified as male reported somewhat higher levels of AI use. These differences are especially noticeable among high-frequency users. Over a third (36%) of men reported using AI professionally at least once per week compared with 30% of women (see Figure 5). Among journalists under 30 years of age, 42% reported using AI at least weekly compared to 29% of those aged 50 and over (see Figure 6).}

Figure 5.

Figure 6. 

Research on the relationship between demographic factors and journalists’ use of AI is still limited. One study of Latin American journalists showed that in some countries, journalists’ perceptions of AI as an opportunity differed significantly by gender (Soto-Sanfiel et al. 2022). However, other studies examining the adoption of innovative technologies in journalism more broadly did not find any gender-specific differences (see, for example, Holman and Perreault 2022).

1.4.2 Level of managerial authority

Our findings show that more senior journalists use AI more frequently for professional purposes. Again, this is particularly evident among high-frequency users. For example, the proportion of journalists in the category with the highest level of managerial authority who use AI daily for journalistic tasks (18%) is three times greater than the proportion in the category with limited editorial responsibility (see Figure 7).

Part of the explanation may be that regular journalists without managerial responsibility have been shown to be worried about the impact of AI on the meaningfulness of their work and their professional identity (Møller et al. 2024). Meanwhile, some of those in more senior positions have been specifically tasked with pushing AI solutions (see, for example, Borchardt et al. 2025). Another explanation could be that the uses being made of AI by those with more management responsibility are subject to fewer limitations than the uses being made by those with less. For instance, the BBC’s AI guidelines are more open to AI being used to provide ‘information, insight or analysis that might aid’ editorial production processes than its use for creating ‘content for publication’ directly (BBC 2025).

Figure 7. 

1.4.3 Reporting beat

AI use was also linked to some of the reporting beats journalists worked on. We found business journalists use AI significantly more frequently than those reporting on lifestyle topics (see Figure 8). For example, 43% of business journalists used AI professionally at least weekly compared with 21% of lifestyle journalists. This is not entirely surprising given the data-heavy and time-sensitive nature of much business reporting. More links between journalists’ beats and their AI use may exist. However, to investigate any such links would require larger samples of journalists working on other beats than were available to us.

Figure 8.

1.4.4 Employment status

Our respondents categorised their employment status in terms of holding a full- or part-time permanent contract, holding a full- or part-time fixed-term contract, being freelance/self-employed, or having an ‘other’ status. Controlling for age and gender, there were no statistically significant associations between employment status and the frequency with which they worked with AI in their journalistic tasks. Although earlier applications of AI in journalism often relied mainly on a level of technological and financial support that only newsrooms could provide, recent developments, including the public launch of ChatGPT in November 2022, have made AI tools widely available, including to freelance journalists.

1.4.5 Media formats produced in

We asked journalists to tell us in which media formats they produced, edited or supervised the production of journalistic content. ^{The number of formats each journalist produces in varies around an average (mean) of 2.41. We found a significant association between the number of formats journalists produce in and how frequently they work with AI professionally. For example, while 48% of journalists producing in one format work with AI professionally at least once a month, 62% of journalists producing in five or more formats do so (see Figure 9). A 2023 survey of UK journalists found that the more media formats journalists produced in ‘the greater their concerns about their emotional and mental well-being’ (Thurman et al. 2025: 24). Journalists may be turning to AI in an attempt to reduce the pressures of producing journalism in multiple formats. Alternatively – or simultaneously – AI may be enabling journalists to produce in more formats. AI-powered tools such as Midjourney, Stable Diffusion, and Wochit have certainly made it easier to produce graphics, cartoons, and illustrations, and to edit video.}

Figure 9.

AI use is not only related to the number of media formats journalists produce in, but also to which specific formats they work with. Because journalists typically produce in multiple formats, the combination of which varies, we explore this using an ordinary least squares (OLS) regression model and plotting the predicted effect of format use on frequency of professional AI use (controlling for age). Being involved in the production of photo-journalism is associated with less frequent use of AI professionally (see Figure 10). This is unsurprising, given the consensus that seems to exist among news audiences and publishers that AI should not be used to generate images that purport to depict real events or situations (see, for example, NYT Trust Team 2024; Simon et al. 2025: 8).

By contrast, being involved in the production of graphics, cartoons, illustrations, or animation is associated with more frequent use. The inherently artificial nature of illustrations, cartoons and animations makes their production with the assistance of AI less problematic in a news context. Text production is also associated with more frequent AI use, as is being involved in video production, which is perhaps a surprise given how, among all UK journalists, the tasks of video generation and video editing are rarely performed with AI assistance (see Figure 2). Nevertheless, it does seem that video production does prompt more frequent AI use. Such use may be compatible with journalistic ethics and audience expectations if such video is clearly illustrative or if AI is used in backend video production process, such as editing (see, for example, Thurman et al. 2024b). Being involved in the production of audio or multimedia stories that use a combination of media formats was not associated with either more or less frequent AI use.

Figure 10.

1.5 Links to job satisfaction

It has often been suggested that the use of AI in journalism will relieve journalists of low-level tasks, freeing up time for them to work on more complex and creative tasks (see, for example, Flew et al. 2012; Lindén 2017; Albizu-Rivas et al. 2024; Noain-Sánchez 2022). Our survey results do not offer support for such suggestions. Indeed, we find that more frequent AI users are more – not less – likely to believe they work on low-level tasks too frequently (see Figure 11). One explanation for this finding could be how AI use is accompanied by new, AI-specific, low-level tasks, such as cleaning data, creating prompts, and checking AI output. Another explanation could be that journalists who feel they work too frequently on low-level tasks are using AI more often to try to lighten this aspect of their workload. Our cross-sectional data cannot reveal whether AI has been beneficial in this regard. Before starting to use AI, some journalists may have felt even less satisfied with the high frequency of low-level tasks they were undertaking. However, our results do show that frequent AI use has not yet relieved journalists of the burden of low-level tasks.

Figure 11.

Our survey also shows that more frequent AI users are not more satisfied with the amount of time they spend working on complex and creative tasks, such as in-depth interviews, interactive storytelling, and investigations. Indeed, those who are most satisfied are journalists who do not use AI at all (see Figure 12). Again, our cross-sectional data cannot reveal the changes AI may have brought about. It is possible that some journalists were less satisfied with the time they spent working on complex and creative tasks before they started using AI. Nevertheless, our results suggest that even if some UK journalists are gaining time from AI-driven efficiencies and investing it in working on complex and creative tasks, the extent of that resource reallocation has not yet transformed this dimension of job satisfaction.

Figure 12.

 

2. What are journalists’ attitudes towards AI?

In this section we explore UK journalists’ attitudes towards AI, specifically whether they see it as an opportunity and/or as a threat, and what they think the potential ethical consequences might be.

Previous research has shown that the UK public tends to be more pessimistic than optimistic. They are slightly more likely to say that generative AI will make their own lives worse rather than better (30% vs 24%), and considerably more likely to think that it will make society worse (49% vs 20%) (Simon et al. 2025). When it comes to journalism specifically, views are perhaps even more pessimistic, with 44% saying that they think generative AI will make their experience of interacting with the news media worse, and only 12% thinking it will make their experience better (Simon et al. 2025).

In this section we show that UK journalists are equally pessimistic. More specifically, AI is seen as a threat to journalism to a much greater extent than it is seen as an opportunity. However, those journalists who regularly use AI for journalistic tasks tend to be relatively more positive about the impact of AI.

2.1 AI as a threat and/or opportunity for journalism

We start by looking at the extent to which UK journalists in our survey think that AI is a threat to journalism and/or an opportunity. As with the UK public, the results point towards pessimism rather than optimism (see Figure 13). A clear majority of UK journalists (62%) think that AI represents a threat to journalism to a large or very large extent, whereas only a small minority of 15% say that it is a large or very large opportunity.

Figure 13.

Of course it is possible to see AI as both a large or very large threat and a large or very large opportunity, and indeed 5% of UK journalists do think this. And it is worth noting that 29% of UK journalists see AI as neither a large threat nor a large opportunity. However, a majority (56%) see AI solely as a large threat, and just 9% see it solely as a large opportunity. The data, then, clearly points to a more pessimistic view of the impact of AI among UK journalists.

This more pessimistic view is held by most demographic groups in our data. Despite younger journalists being more likely to use AI (see Section 1.4.1), there are no real differences by age group on whether AI represents a large opportunity or a large threat (see Figure 14). Male journalists (17%) are slightly more likely than women (12%) to see AI as a large opportunity for journalism, but both are much more likely to see it as a large threat (62% and 61% respectively).

Figure 14.

Differences by the media cultural background of the outlet journalists primarily work for are also small (see Figure 15). Newspaper (11%) and magazine journalists (16%) are slightly less likely to see AI as a large opportunity, as compared to broadcast (television and radio) journalists (22%), and newspaper journalists are also slightly more likely to see AI as a large threat than those that work for the other types of publisher. Journalists that work for internet-native publishers sit somewhere in between print and broadcast journalists.

Figure 15.

It is only when we consider UK journalists’ employment in a bit more detail that larger differences start to emerge. While the number of years someone has worked as a journalist seems to be unrelated to their attitudes towards AI, there are some significant differences by professional seniority (see Figure 16). Journalists with limited editorial responsibility tend to be quite pessimistic about the effects of AI. Mid-level journalists with operational authority within a specific division or department are slightly less likely to see AI as a threat and slightly more likely to see it as an opportunity. However, the real difference is between the most senior group – with strategic authority to make long-term decisions about the media organisation – and the rest. Around one third (34%) of those in this group see AI as a large opportunity and around half (53%) see it as a large threat. As well as being more likely to use AI in their professional life (see Section 1.4.2), journalists in this group are often the driving force behind the adoption and integration of AI in newsrooms, so it is little surprise that they see it more positively, although it is perhaps a surprise that even among this group the data still skews negative rather than positive.

Figure 16.

There are larger differences still when we look at journalists with different levels of AI knowledge and usage (see Figure 17). Journalists with relatively high levels of AI knowledge (those who correctly answered at least five of the six factual true or false AI questions we asked them in the survey) are more likely to see AI as a large opportunity (24%), although the proportion that see it as a large threat is roughly similar to those in the medium- and low-knowledge groups (three or four correct answers and fewer than three correct answers, respectively).

The largest differences can be found when we split the data by how frequently the respondents use AI for journalistic tasks. ^{Journalists that use AI less than once a week (including those that never use it) tend to be more pessimistic than the average UK journalist. However, those that use it on a weekly basis, and especially those that use it daily, are more likely to see it as a large opportunity (25% and 45% do so, respectively), and less likely to see it as a large threat. In fact, those that use AI daily for journalistic tasks are as likely to see AI as a large opportunity (45% do so) as a large threat (48%) – one of the few groups in the data to not have an overwhelmingly pessimistic view of the potential impact of AI. This highlights the importance of technology use for shaping attitudes.}

Figure 17.

2.2 UK journalists’ concerns about the potential ethical consequences of AI in journalism

Part of the perceived threat to journalism posed by AI is the potential ethical concerns it raises. Even if they acknowledge that AI can perform certain journalistic tasks to a high standard, some worry that there may be indirect ethical consequences of deploying AI in the newsroom.

To dig deeper into this issue we used the survey to ask about UK journalists’ level of concern over a range of different potential consequences, with concerns covering the potential negative impact on journalistic values, content, public attitudes, and privacy.

Overall levels of concern about these potential negative consequences among UK journalists are very high (see Figure 18). More than half say that they are extremely concerned about the negative impact on public trust in journalism (60%), the journalistic value of accuracy (57%), and the originality of journalistic content (54%), for example. Half or more are extremely concerned or strongly concerned about all the other consequences asked about, ranging from concern over the production of intransparent journalistic content (69% are extremely or strongly concerned) to inadvertent exposure of personal data (51%), which drew least widespread concern from our survey sample.

Figure 18.

It is worth noting that some of these concerns are currently still hypothetical, no matter how plausible they may seem. It is not yet clear what the effect of the impact of AI will be on public trust, especially in the medium to long term. Other concerns, however, do already have some empirical support. For example, in early 2025 Apple suspended its AI-generated news alert service after inaccurate mobile alerts branded with the BBC’s logo were sent to users.

In line with the data showing that most UK journalists across most demographic and employment groupings perceive AI as a large threat to journalism, most are also strongly or extremely concerned about each potential ethical consequence. In most cases the differences between groups are small. Taking age as an example (see Figure 19), we typically see quite small differences between the youngest and oldest age groups, both in their views on individual ethical concerns and if we average across all seven concerns. There are some exceptions, however, such as the 12 percentage point gap between the proportion of 18- to 34-year-olds and those aged 55 and over that are strongly or extremely concerned about the production of intransparent journalistic content. However, it is not possible here to fully explore what might be behind such differences.

Figure 19.

Average gaps in concern by gender, years worked as a journalist, media cultural background and seniority are all equally small, at 5 percentage points or less. However, there are some differences by variables specifically related to AI. As we can see in Figures 20 and 21, concern tends to be higher among those with higher levels of AI knowledge, but considerably lower among those that use AI for journalistic tasks on a daily basis.

Figure 20.

Figure 21.

As a recent systematic review shows (Oh and Jung 2025), researchers have used case studies and interviews to map the range of different threats and opportunities associated with AI as perceived by journalists. However, the strength of the data we offer here is that it shows how widely (or not) threats and opportunities are perceived – and how they compare. When we do this, we see that the threat of AI (and related concerns over potential ethical consequences) are more widely perceived than opportunities – although it is noteworthy that those who regularly use AI for journalistic tasks tend to be relatively more positive about the impact of AI. As the use of AI grows within newsrooms, does that mean that journalists will come to be more positive about its impact over time?

3. Which news organisations use AI?

AI can be applied to a wide range of individual journalistic tasks, and it is clear from Section 1 that a significant number of UK journalists are now using AI in their work. However, this data does not tell us about the extent to which AI has been integrated into newsroom processes. This section explores AI integration in UK newsrooms and how this varies by different outlet types.

Leaving aside those newsrooms that do not currently use AI in any capacity, we can imagine different levels of AI integration, ranging from limited integration, where AI is minimally integrated into some specific tasks or processes, through to full integration, where AI is fully integrated into all aspects of newsroom processes.

In our survey we asked journalists about the extent AI was integrated into the newsroom processes at their main employer’s news outlet. In this section we will show that so far the majority of UK newsrooms have only achieved limited integration, or have not integrated AI at all, although there are some differences by the type of news outlet.

Before exploring this data in more detail it is important to point out that some UK journalists, depending on factors including where they are situated within their organisation and what their working conditions are, are likely unaware of how integrated AI is. This is likely why a significant minority – one fifth (20%) – of UK journalists answered ‘don’t know’ to this question. Because this is a large proportion, and because uncertainty is less meaningful on this question, we exclude the ‘don’t know’ responses from the analyses in this section.

3.1 The level of AI integration in newsrooms

Four out of ten respondents (40%) report that AI has not been integrated into the newsroom processes at their main employer’s outlet at all, meaning 60% of respondents work in newsrooms that have integrated AI to some extent (see Figure 22). ^{However, a large majority of respondents who say that AI has been integrated described that integration as limited (three quarters of the 60% whose newsroom has integrated AI). Just 11% said that the level of AI integration was moderate, with very few describing it as extensive (3%). Only 1% said ‘AI technology is fully integrated into all aspects of newsroom processes at my main news outlet’.}

Figure 22.

3.2. How AI integration varies by news outlet

This picture stays broadly the same if we break the data down by the media cultural background of the main outlet that the journalist works for (see Figure 23). Those that work for newspaper and internet-native outlets are more likely to report moderate levels of AI integration (and are thus less likely to report no integration) than those who primarily work for broadcasters and magazines. However, hardly any journalists report extensive or full integration, regardless of the type of outlet they mainly work for.

Figure 23.

Journalists working for private or commercial media are more likely to report limited AI integration (as opposed to none) than those working for publicly owned media (see Figure 24), and those working for conglomerates are more likely to report moderate levels of AI integration than those working for independent outlets that are not part of a larger group (see Figure 25). Although independent outlets may be more flexible and can therefore allow their journalists to adopt AI for certain journalistic tasks in an ad hoc way, conglomerates are more likely to be able to integrate AI through the rollout of tools and systems companywide, in part due to larger resources and dedicated AI personnel.

Figure 24.

Figure 25.

Although we do see some small differences in AI integration by outlet type in the data, it is also clear that, overall, UK newsrooms have only so far achieved limited integration, or have not integrated AI at all. However, it is important to point out that UK journalists think that this is likely to change. If we look at the percentage point difference between those that say their main news outlet plans to increase AI integration and those that say it plans to decrease AI integration, journalists overwhelmingly say they think integration will increase (63 percentage points) (see Figure 26). The belief that AI integration will increase is even stronger among journalists whose main outlet is a broadcaster (+82), a conglomerate (+77) or is publicly owned (+85). Around 11% on average said that they expect their main news outlet to maintain their current level of AI integration.

Figure 26.

Although AI integration is often described as limited, the data suggests that this may be partly due to inertia that can make big changes difficult. Overall, UK journalists are around twice as likely to describe their main outlet’s stance on AI to be either supportive or strongly supportive (39%) than either opposed or strongly opposed (20%) (see Figure 27).

Figure 27.

Although there are some small differences in support between different outlet types, in each case the proportion of UK journalists that describe their main employer as supportive is always larger than the proportion that see their employer’s stance as opposed. Journalists working for internet natives (46%) are more likely to describe them as being supportive than those working for newspapers (38%), magazines (32%), or broadcasters (28%).

Overall, then, the data suggests that while AI integration in newsrooms is currently limited, UK journalists see their employers as being relatively positive about AI, and they strongly expect AI integration to increase in the future.

4. What are news organisations’ approaches towards AI?

The integration of AI into newsrooms raises a number of related practical and organisational issues that news outlets must consider. In this section we explore whether UK news outlets have established AI guidelines and protocols, what types of AI tools they use, and whether or not they provide AI training to their staff.

We show that most UK journalists say that their main news outlet has at least some guidelines or protocols in place for the use of AI, but most say AI training is not provided. Most journalists say their main outlet only uses third-party AI tools for their journalistic work. However, there are differences between outlets, with publicly owned outlets and conglomerates – which are disproportionately newspapers and broadcasters – more likely to have guidelines, training, and in-house tools.

4.1 Protocols and guidelines around AI use

We begin by considering protocols or ethical guidelines to address a range of issues raised by the use of AI. ^{Around 40% of UK journalists report that their main news outlet has established protocols or ethical guidelines for most of the issues we asked about, including human oversight and control (44%), data privacy and security (43%), transparency (42%), application areas (what AI can be used for) (41%), and accountability/responsibility (40%) (see Figure 28). However, only 27% say that their main news outlet has protocols or guidelines around bias and fairness. This could be because certain biases are baked into generative AI models because of how they are trained, and there may be less that news organisations can do to address this. Overall, 60% of UK journalists say that their main news outlet has a protocol or guideline for at least one of these issues.}

Figure 28.

There are large differences in the establishment of protocols and guidelines by types of news outlet, particularly concerning their ownership. Specifically, journalists whose main news outlet is publicly owned are more likely to report that their outlet has established protocols and guidelines for all of the issues we asked about (see Figure 29). Whereas around 40% of those working for private/commercial media say their outlet has guidelines on most issues (although 23% for guidelines around bias and fairness), around 70% of those working for publicly owned media say the same. This could be because publicly owned media can expect more scrutiny about their use of AI in the newsroom, and are often expected to uphold higher standards associated with public service media. It is also noteworthy that although publicly owned outlets are more likely to have bias and fairness guidelines than private/commercial media, it is still the issue least likely to have a formal policy.

Figure 29.

Related to this, most UK journalists report that ethical guidelines are updated as needed (79%), rather than at regular fixed intervals (18%) (see Figure 30). This suggests that news outlets are typically adopting a flexible approach to their AI policies that adapts to new issues as they emerge.

Figure 30.

4.2 Provision of AI training

Around one third (32%) of UK journalists report that their news organisation provides AI training (see Figure 31). It should be kept in mind, however, that many journalists are using AI infrequently or not at all, and that most newsrooms have only integrated AI to a limited extent. Journalists in newsrooms with more extensive integration are more likely to say their news organisation provides AI training.

UK journalists whose main outlet is publicly owned are more likely to say their news organisation provides AI training (57%) than those working for private/commercial media (29%). This is likely also because of the increased scrutiny and higher expectations associated with public service media referred to in Section 4.1 on ethical guidelines and protocols. Journalists working for conglomerates are also more likely to say their news organisation provides AI training (50%) than those working for independent outlets (14%) – likely reflecting their increased resources and ability to roll out training programmes companywide. Outlets with a broadcast or newspaper background are more likely to be publicly owned or part of a conglomerate than internet natives or magazines, which could explain why those working for them are also more likely to say that their news organisation provides AI training.

Figure 31.

4.3 Types of AI tools

Finally, we look at the types of AI tools used by UK journalists’ main news outlets. In the survey we made a distinction between tools developed in-house and tools developed by third parties (see Figure 32). Given the nature of the technology and the skills and resources required to develop it in-house, it is perhaps little surprise that only 9% of journalists say that their main outlet only uses in-house tools. Over half (57%) say that their main news outlet only uses third-party tools, with 34% saying their outlet uses a combination of both third-party and in-house tools.

Figure 32.

Those working for conglomerates and publicly owned media are a little more likely to say their main outlet only uses in-house tools, and considerably more likely to say it uses both in-house and third-party tools, making this mixed approach the dominant one among conglomerates (47%) and publicly owned media (59%). Again, this helps explain why we found that newspapers and broadcasters are also more likely to combine the use of in-house tools with third-party tools. Across all areas covered in this section – guidelines and protocols, the provision of training, and the use of different types of tools – we see differences between newspapers and broadcasters on the one hand, and magazines and internet natives on the other. However, these differences are more likely rooted in ownership and size rather than media cultural background.

5. Surveying a representative sample of UK journalists: Methodology

This report is based on an online survey carried out between 29 August and 4 November 2024. Two of the authors of this report (Neil Thurman and Sina Thäsler-Kordonouri, hereafter ‘the team’) oversaw the sampling and the data collection, processing, and cleaning. The other author and the Reuters Institute were not involved in the sampling, survey design, data collection, data processing, or data cleaning, but carried out secondary analysis of the final anonymised dataset.

After data cleaning, the survey had 1,004 responses, a sample that is broadly representative of the total population of UK journalists. ^{In this section we lay out the methodology in detail and assess the representativeness of our sample.}

5.1 The questionnaire

The survey contained questions about journalists’:

• Personal characteristics, specifically: gender, age, openness to experiences in general (adapted from Soto and John 2017), and AI knowledge (adapted from Soto-Sanfiel et al. 2024).
• Employment conditions, specifically: years of work experience, employment contract, management responsibility, satisfaction with time spent on low-level tasks (such as processing and cleaning data), satisfaction with time spent on complex and creative tasks (such as in-depth interviews), perceptions of job security (adapted from Vander Elst et al. 2013), workplace stress (adapted from The American Institute of Stress 2009), and level of editorial freedom.
• Working routines, specifically: beat worked on; media formats worked with (text, photographs etc.); frequency of journalistic use of AI in general; and frequency of use of AI for 31 specific journalistic tasks (e.g. translation, still image generation etc.).
• Perceptions of: the extent to which AI is a threat to journalism; the extent to which AI is an opportunity for journalism; and concern about seven possible ethical consequences of AI use in journalism.
• Main employer (if any), specifically the employer’s: media cultural background (e.g. newspaper, internet native); level of newsroom AI integration; use of AI for specific newsroom tasks; use of in-house and/or third-party AI tools; stance on AI; future plans for AI integration; provision of training on AI; provision of AI protocols; size (number of employees, number of news rooms); ownership (e.g. public, private); and membership of a larger media conglomerate (if at all).

The full questionnaire is available on Figshare (Thurman and Thäsler-Kordonouri 2025).

The questions in italics were adapted from the UK Worlds of Journalism Study Wave 3 questionnaire (Lauerer et al. 2025; Thurman et al. 2024a).

5.2 Sampling strategy

In order to build our sample, we first obtained a list of journalists’ names, email addresses, and professional affiliations from the Roxhill database of journalists and media outlets. In the Roxhill database, journalists are associated with particular outlet types, such as ‘National (newspapers)’. Contact details were downloaded for UK-based journalists working across all of Roxhill’s outlet types, specifically: blogs; business ‘trade’ magazines; consumer magazines; national and regional newspapers; news and picture agencies and newswires; podcasts; national and regional radio stations; national and regional TV stations; and ‘Freelance’ (also a Roxhill outlet type).

Because we wanted to include UK foreign correspondents – journalists working for UK publications overseas – we also downloaded from Roxhill the contact details of journalists based outside the UK. To eliminate journalists from this list who did not work for UK-based publications, a list of all UK media outlets was downloaded from Roxhill (N=12,831). Only journalists who worked for at least one of these UK media outlets were retained.

With each of our lists of UK-based journalists and UK foreign correspondents, we removed contacts that had no email address and attempted to remove all those that were for generic desks (e.g. news desk, foreign desk, sports desk). Because some journalists were associated with more than one outlet type – for example, national radio and national television – a deduplication process was also undertaken. This resulted in a list of 35,775 UK-based journalists and a list of 4,463 UK foreign correspondents. These two lists were combined and after some duplicates and further contacts with generic email addresses were removed we had a final list of 40,040 contacts.

There is no official record of the population of journalists in the UK (or journalists working for UK-based publications outside the UK). However, at the time of writing in June 2025, data was available from the 2021 Census for England, Wales and Northern Ireland on the numbers of people living in England, Wales and Northern Ireland who declared they were employed as editors, journalists, and reporters, using these two 2020 Standard Occupational Classifications:

SOC 2491: Newspaper, Periodical and Broadcast Editors

SOC 2492: Newspaper, Periodical and Broadcast Journalists and Reporters

The 2021 Census estimates 30,060 people worked as newspaper, periodical and broadcast editors in England and Wales in March–May 2021 (ONS 2023). Another 24,630 people were estimated to be working as newspaper, periodical and broadcast journalists and reporters, giving a total of 54,690. The 2021 Census also estimated that there were 674 newspaper and periodical journalists and reporters and 324 newspaper, periodical and broadcast editors in Northern Ireland (NISRA 2023). That makes a total of 55,688 for England, Wales and Northern Ireland.

Because labour market data from the 2022 Scottish Census were not available at the time of writing, we had to estimate the number of journalists in Scotland. The UK’s Office for National Statistics’ Labour Force Survey (ONS 2021) showed that, in February, March and April 2021 (around the time the 2021 Census was taken) the number of people aged 16 and over in employment for the countries in the United Kingdom was as follows:

• England and Wales: 28,712,314
• Northern Ireland: 823,215
• Scotland: 2,638,571

Therefore, approximately 0.19% of the employed population of England and Wales are journalists and approximately 0.12% of the employed population of Northern Ireland. If we take the higher proportion for Scotland (0.19%), we can estimate that 5,026 journalists work in Scotland. That would bring the estimated total number of journalists in the UK to 60,714 (but excluding foreign correspondents working for UK publications). This means that our list of journalists (based in the UK) generated by the Roxhill database represents around 59% of the total population in the UK.

If the list of 4,463 foreign correspondents working for UK publications generated by the Roxhill database also represents around 59% of the total population, then the total population of foreign correspondents would be 7,565. Adding that to the estimated total number of journalists in the UK (60,714) equals 68,279.

Because we wanted our sample size to have a confidence level of at least 95% and a maximum error margin of 3%, with an estimated population of 68,279 we aimed for a sample size of at least 1,051. Based on our experience of previous surveys of UK journalists (Thurman et al. 2016; Thurman et al. 2025), we expected the response rate to our online survey to be relatively low for this hard-to-reach group. As a result, we decided to send email invitations to a random selection of 18,263 journalists from our list of 40,040.

Including reminders, journalists received up to ten email invitations to participate in the survey between 29 August and 31 October 2024. Participation was by invitation only, and the survey was closed on 4 November 2024. The survey was hosted on the Qualtrics online survey platform. The project received ethical approval from the Institutional Review Board of the Department of Media and Communication at LMU Munich on 24 July 2024.

5.3 Exclusions

We conducted extensive research – including via LinkedIn, Twitter/X, the Roxhill Media database, Cision One, and news outlets featuring the journalists’ work – to be sure each respondent met our definition of a journalist. Respondents were excluded if:

• they did not work for a media outlet with an identifiable focus on providing news
• they did not work for a news outlet that had a UK base and was aimed, at least in part, at a UK audience, and
• they did not earn at least 50% of their income from journalism or work for at least 18 hours per week as a journalist (which is 50% of a regular working week in the UK).

Furthermore, respondents were excluded if they did not complete the survey up to and including the question asking them ‘How frequently do you personally work with AI in your journalistic tasks?’, a question that appeared approximately 47% of the way into the survey, and if they had ‘straightlined’ questions for which that response pattern was contradictory.

After these exclusions were made, our final sample was 1,004. If we take the total population of UK journalists to be 68,279 (see above), our survey’s sample size can be considered to be robust by the standards of social survey research, with a maximum margin of error of 3% at a confidence level of 95% for some questions. The margin of error was higher – up to 6% – for some other questions that were not answered by all respondents. ^{As such, throughout the report, we do not consider differences of +/– 3 percentage points or lower to be meaningful, and when analysing small subgroups, we often set this cut-off even higher.}

5.4 Sample description

The average age of the journalists in our final sample was 47.1 (SD = 12.39). Slightly more (55.3%) identified as male than as female (44.5%). Only 0.2% stated their gender as ‘other’. Respondents had worked as journalists for an average of 21.8 years (SD = 11.88). Just over half worked full time on permanent contracts (55.1%). A little more than a third worked freelance or were self-employed (37.3%).

Regarding seniority, 39.1% of our sample said they had limited editorial responsibility, while another 44.4% said they had day-to-day departmental operational authority. The rest (16.5%) reported having long-term strategic authority across newsrooms or divisions.

Regarding professional specialisation, over half of the journalists reported working on a specific beat (58.4%), with the largest proportion of those specialising in business issues, followed by lifestyle specialists and those on the culture beat.

Journalists were asked about the media cultural background – from newspaper to news agency – of their main news outlet. Excluding the 14.9% who did not work for one main outlet, for 35.6% their main outlet had a magazine background; internet-native 23.4%; newspaper 22.4%; radio 7.3%; TV 7.1%; and news agency 2.9%. Just over 1% of journalists described their main outlet as having an ‘other’ background.

Of the journalists who worked for one main news outlet and knew its ownership, 80.3% worked for a commercially owned outlet, 14.7% for a publicly owned one, and 5% for an outlet that either had community, state, or some other form of ownership.

Of the journalists who worked for one main news outlet, about half (52.6%) worked for an outlet that was part of a larger media conglomerate.

5.5 Data cleaning and anonymisation

Each response to each question was examined and some changes were made. For example, a respondent who classified their beat as ‘other’ was reclassified to one of the items in the predefined list of beats (9: Lifestyle [e.g. Food & Drink, Fashion, Beauty, Travel, Home & Garden, Health’]) because they had written that their beat was ‘travel’. After data cleaning, an anonymised version of the data was shared with the Reuters Institute and the other author of this report.

5.6 Response rate and representativeness

Although we sent email invitations to a random selection of 18,263 journalists from our list of 40,040, not all invitations were received. Invitation emails were sometimes rejected as spam, or journalists never received them because they were away on holiday. It was common for between a quarter and a third of the emails sent in each distribution to fail or bounce. If we assume that all journalists whose email inboxes rejected at least one invitation did successfully receive an invitation on another occasion, our response rate (calculated with reference to the journalists we kept in the final sample) would be 5.5%. However, this assumption is implausible because spam policies are unlikely to reject one email and allow another, some journalists (for example those on parental leave) were out of the office for the whole of the two months of the survey fieldwork, and some email addresses were out of date because journalists had moved employers. Therefore, we believe our actual response rate to be higher than 5.5%, although we cannot say by how much.

Although there is no central, all-inclusive list of journalists in the UK (NCTJ 2012: 12), we have used data from the 2021 Census of England, Wales and Northern Ireland on the population of journalists in England and Wales to help judge the representativeness of our sample of UK-based journalists. However, in interpreting the comparisons we have made, it is important to bear in mind the following differences: our sample includes journalists living in Scotland and Northern Ireland; we excluded journalists who did not earn at least 50% of their income from journalism or work for at least 18 hours per week as a journalist; we also excluded journalists who worked for outlets that we deemed did not have an identifiable focus on providing news or did not have a UK base and were aimed, at least in part, at a UK audience. The 2021 Census classified individuals as journalists solely on the basis of their self-identification as such.

Keeping these limitations in mind, and the differences in the timing of the respective fieldwork, as shown in Table 1, our sample does have a higher proportion (by about 7 percentage points) of men than the journalists recorded as male at birth in the 2021 England and Wales Census data. This male skew may be the result of a response bias due to men’s apparent higher willingness to use AI (Gnambs et al. 2025), higher use of generative AI (Otis et al. 2024), and more positive attitudes towards the technology (Bergdahl et al. 2023).

Our sample is also older, with the proportion over 39 years of age about 17 percentage points higher than in the England and Wales Census. This difference could, in part, be due to the minimum requirement that we set regarding journalists’ income/working hours, which may have excluded some younger, part-time journalists included in the 2021 Census data.

Although respondents in our sample skew older and somewhat more male than those who identified as journalists in the 2021 Census of England and Wales, we believe that this difference is not an indication of any fundamental flaw in our sampling strategy but rather, as discussed above, the result of differences in the respective inclusion criteria used as well as a small topic-related response bias.

Table 1.

References

• Albizu-Rivas, I., Parratt-Fernández, S., Mera-Fernández, M. 2024. ‘Artificial Intelligence in Slow Journalism: Journalists’ Uses, Perceptions, and Attitudes’, Journalism and Media 5(4), 1836–50. https://doi.org/10.3390/journalmedia5040111
• American Institute of Stress. 2009. ‘The Workplace Stress Scale: Attitudes in the American Workplace VII.’
• BBC. 2025. ‘Guidance: The Use of Artificial Intelligence’, 24 June.
• Bergdahl, J., Latikka, R., Celuch, M., Savolainen, I., Mantere, E.S., Savela, N., Oksanen, A., 2023. ‘Self-Determination and Attitudes Toward Artificial Intelligence: Cross-National and Longitudinal Perspectives’, Telematics and Informatics, 82, 102013.
• Borchardt, A., Mulhall, E., Bremme, K., López Garrido, B., Johanny, Y. 2025. Leading Newsrooms in the Age of Generative AI.
• Cools, H., Diakopoulos, N. 2024. ‘Uses of Generative AI in the Newsroom: Mapping Journalists’ Perceptions of Perils and Possibilities’, Journalism Practice, 1–19. https://doi.org/10.1080/17512786.2024.2394558
• Flew, T., Spurgeon, C., Daniel, A., Swift, A. 2012. ‘The Promise of Computational Journalism’, Journalism Practice 6(2), 157–71. https://doi.org/10.1080/17512786.2011.616655
• Gnambs, T., Stein, J.P., Zinn, S., Griese, F., Appel, M. 2025. ‘Attitudes, Experiences, and Usage Intentions of Artificial Intelligence: A Population Study in Germany’, Telematics and Informatics 98, 102265. https://doi.org/10.1016/j.tele.2025.102265
• Holman, L., Perreault, G. P. 2022. ‘Diffusion of Innovations in Digital Journalism: Technology, Roles, and Gender in Modern Newsrooms’, Journalism 24(5), 938–57. https://doi.org/10.1177/14648849211073441
• Kalai, A. T., Nachum, O., Vempala, S. S., Zhang, E. 2025. ‘Why Language Models Hallucinate.’ https://doi.org/10.48550/arXiv.2509.04664
• Lauerer, C., Grünewald, M., Hanitzsch, T., Godole, J., Zguri, R., Godole, B. … Alakoaa, A. 2025. Worlds of Journalism Study Wave 3 2021–2024 Consolidated Dataset [Data set]. Zenodo. https://doi.org/10.5281/zenodo.15280739
• Lindén, T. C. G. 2017. ‘Algorithms for Journalism: The Future of News Work’, The Journal of Media Innovations 4(1), 60–76. https://doi.org/10.5617/jmi.v4i1.2420
• Møller, L. A., van Dalen, A., Skovsgaard, M. 2024. ‘A Little of that Human Touch: How Regular Journalists Redefine Their Expertise in the Face of Artificial Intelligence’, Journalism Studies, 26(1), 84–100. https://doi.org/10.1080/1461670X.2024.2412212
• NYT Trust Team. 2024. ‘How The New York Times Uses A.I. for Journalism’.
• NCTJ. 2012. Journalists at Work. Saffron Walden: National Council for the Training of Journalists.
• NISRA. 2023. ‘Census 2021 Main Statistics Labour Market Tables’, Northern Ireland Statistics and Research Agency, 21 March
• Noain-Sánchez, A. 2022. ‘Addressing the Impact of Artificial Intelligence on Journalism: The Perception of Experts, Journalists and Academics’, Communication & Society 35(3), 105–121. https://doi.org/10.15581/003.35.3.105-121
• Oh, S., Jung, J. 2025. ‘Harmonizing Traditional Journalistic Values With Emerging AI Technologies: A Systematic Review of Journalists’ Perception’, Media and Communication 13, 9495. https://doi.org/10.17645/mac.9495
• ONS. 2021. ‘Labour Force Survey – National and Regional – Headline Indicators (Seasonally Adjusted)’, Office for National Statistics
• ONS. 2023. ‘Occupations of Those in Employment, by Age and Sex, England and Wales, Census 2021’, Office for National Statistics, 31 May.
• Otis, N.G., Delecourt, S., Cranney, K., Koning, R. 2024. Global Evidence on Gender Gaps and Generative AI. Harvard Business School Working Paper, No. 25-023. (Revised August 2025.)
• Simon, F., Nielsen, R. K., Fletcher, R. 2025. Generative AI and News Report 2025: How People Think About AI’s Role in Journalism and Society. Oxford: Reuters Institute for the Study of Journalism. https://doi.org/10.60625/risj-5bjv-yt69
• Soto, C. J., John, O. P. 2017. ‘The Next Big Five Inventory (BFI-2): Developing and Assessing a Hierarchical Model with 15 Facets to Enhance Bandwidth, Fidelity, and Predictive Power’, Journal of Personality and Social Psychology 113(1), 117–43. https://doi.org/10.1037/pspp0000096
• Soto-Sanfiel, M. T., Ibiti, A., Machado, M., Marín Ochoa, B. E., Mendoza Michilot, M., Rosell Arce, C. G., Angulo-Brunet, A. 2022. ‘In Search of the Global South: Assessing Attitudes of Latin American Journalists to Artificial Intelligence in Journalism’, Journalism Studies 23(10), 1197–1224. https://doi.org/10.1080/1461670X.2022.2075786
• Soto-Sanfiel, M. T., Angulo-Brunet, A., Lutz, C. 2024. ‘The Scale of Artificial Intelligence Literacy for All (SAIL4ALL): A Tool for Assessing Knowledge on Artificial Intelligence in All Adult Populations and Settings’, SocArXiv. April 30. https://osf.io/preprints/socarxiv/bvyku_v1
• Thurman, N., Henkel, I., Thäsler-Kordonouri, S., Fletcher, R. 2025. UK Journalists in the 2020s: Who They Are, How They Work, and What They Think. Oxford: Reuters Institute for the Study of Journalism.
• Thurman, N., Henkel, I., Thäsler-Kordonouri, S. 2024a. ‘UK Version of the Worlds of Journalism Study Survey (3rd Wave)’, Figshare, 6 June. https://doi.org/10.6084/m9.figshare.25983157
• Thurman, N., Stares, S., Koliska, M. 2024b. ‘Audience evaluations of News Videos Made with Various Levels of Automation: A Population-Based Survey Experiment’, Journalism 26(1), 3 –23. https://doi.org/10.1177/14648849241243189 (Original work published 2025)
• Thurman, N., Thäsler-Kordonouri, S. 2025. ‘Quantitative Survey Questionnaire on Artificial Intelligence (AI) in Journalism Fielded to Journalists’, Figshare, 11 June. https://doi.org/10.6084/m9.figshare.29251091
• Thurman, N., Kunert, J., Cornia, A. 2016. Journalists in the UK. Oxford: Reuters Institute for the Study of Journalism. https://doi.org/ 10.60625/risj-x53v-0w57
• Vander Elst, T., De Witte, H., De Cuyper, N. 2013. ‘The Job Insecurity Scale: A Psychometric Evaluation Across Five European Countries’, European Journal of Work and Organizational Psychology 23(3), 364–80. https://doi.org/10.1080/1359432X.2012.745989

---