78% of UK professionals now use generative AI at work

New CLA research reveals how generative AI is reshaping content behaviour across UK workplaces.

Generative AI has fundamentally transformed content behaviour across UK workplaces, according to new research published by the Copyright Licensing Agency (CLA). The 2026 Business Behaviour Research, based on a survey of 1,000 UK professionals, found that 78% now use some form of generative AI at work, up from 2 in 5 in 2024, with AI now integrated into 3 in 4 workplaces.

UK professionals continue to rely heavily on published content for day-to-day business activities, from training and development to industry research. But the rapid adoption of generative AI has introduced entirely new ways of creating, copying and sharing content, and with that, new challenges that organisations cannot afford to ignore.

Access the report

A new content landscape

Perhaps the most striking finding is the rise of AI-generated content as a workplace source. In 2024, output from generative AI was the least selected content source in the survey, behind even audio. In 2026, it has jumped to the second most used source overall (38.1%), behind only government information and statistics (40.5%). Traditional sources such as videos, online photos, industry publications and news media have remained broadly stable, meaning AI has joined the content mix rather than replacing it.

This creates a more complex picture for organisations trying to manage how content is used. Professionals are now drawing on a wider blend of material, and tracking what has been used, where it came from and whether permission is needed is becoming harder.

The feedback loop

The research reveals a pattern of behaviour that should concern any organisation serious about managing risk. AI outputs are now the most common input fed back into AI tools, cited by 34.7% of users, creating a feedback loop where content is continuously reused and recycled as the basis for new prompts, amplifying copying behaviour in the process.

This creates a feedback loop where content is continuously reused and recycled. Crucially, traditional content sources such as news media, books, journals and industry publications are still being used at the same levels as in 2024. That content does not disappear when it is used to inform an AI output. It is still in there. But when that output is then fed back into another tool as a prompt, the trail back to the original third-party content fades.

The concern is one of visibility. With each cycle of reuse, it becomes harder for organisations to know what copyright protected material is embedded in the content their teams are producing and sharing, and harder still to check whether proper permissions are in place.

Copying at speed

From using AI outputs as prompts for other tools to copying and pasting generated text into presentations, 4 in 5 generative AI users copy from it at least once a week, the highest rate of any content source. AI outputs are almost twice as likely as other content to be copied by being used as a prompt or input into another AI tool.

Once copied, content moves fast. Email is the most common sharing method (54.6%), followed by messaging platforms like Teams (42%) and WhatsApp (28.8%). The speed and reach of digital sharing means copied content can travel across an organisation in seconds, often far beyond its original context.

Confident, but less cautious

The research highlights a growing disconnect between confidence and behaviour. Most organisations have IP policies in place (69.6%) and most professionals rate their understanding of copyright as at least reasonable (58.6%), which on the surface seems reassuring.

But the ease and speed of AI-assisted content creation has normalised copying behaviour, with fewer professionals viewing copyright infringement as a genuine risk to their organisation. Compared to 2024, fewer say they would seek permission before copying (down 5 points) and fewer agree that infringement is a business risk (down 5.2 points). More than half (54%) say they would not hesitate to copy or share content if work demanded it.

Five steps an organisation can take

CLA’s research points to a clear conclusion: the speed of AI adoption is outpacing the policies designed to manage it. Organisations can take five practical steps: establish clear AI usage policies, define what content can be used as inputs into AI tools with proper licensing in place, invest in employee training on copyright and responsible AI use, evaluate the security and data practices of AI tools on an ongoing basis, and monitor usage to identify risks early.

Navigating this new environment while ensuring copyright compliance and appropriate licensing is more important than ever. The organisations that get ahead will be those that embrace generative AI with the right framework in place, enabling their people to innovate with confidence.

Access the report

What CLA is doing in the generative AI space

CLA is not just highlighting the challenges. It is actively building solutions to help organisations and creators navigate the changing landscape with confidence.

• Workplace Generative AI Permissions: CLA has extended its business and some public sector licences to include permissions for the copying and inclusion of published content to prompt permitted generative AI tools to generate outputs at work, giving organisations legal certainty while ensuring creators are fairly remunerated. Find out more
• Generative AI Solution: CLA has developed a collective licensing solution that will provide AI developers with a clear, lawful pathway to access published content for training and developing AI models, while ensuring publishers and authors are renumerated. Find out more