UK firms risk breaching strict copyright laws as employees admit to regularly using third-party content in AI prompts, despite knowing the risks.

That’s according to a new paper from the Copyright Licensing Agency (CLA), The GAI Revolution, which found that 82% of workers who use genAI tools are guilty of incorporating protected, third-party content into their prompts at work.

The CLA’s report, based on a survey of nearly 4,000 UK professionals, identifies a disconnect between copyright awareness and responsible genAI use in the workplace. Despite more than half of UK employees now using genAI tools at least weekly, only 24% reported a strong understanding of the copyright terms and conditions of these tools.

Added to that, although more than three-quarters (77%) of employees acknowledged these risks, over half remain unconcerned about the potential implications of including copyrighted content in genAI inputs.

While 61% believe their organisation has a policy regarding the use of genAI at work, only 38% think that policy covers copyright issues, and although 76% agreed that AI users should seek permission before using copyright protected materials in prompts, the CLA found that this is muted in practice.

Almost half (45%) of workers said they’re not worried about potential copyright issues connected to their AI use at work, with just 38% expressing any concerns about including third-party material in AI prompts.

Among the most common copyright-protected materials used in AI inputs are images or photos (64%), news articles (55%), business papers (53%), journal articles (48%), and books or ebooks (46%).

The outputs generated from these materials are then used in emails (60%), reports (56%), articles (46%), social media posts (40%), and even as the basis for policies (21%), all of which the CLA found to be shared both in-house and externally.

With more than 70% of UK professionals expected to be using genAI for work by the end of this year, the CLA said that there is a significantly increased risk of unintentional copyright infringement with third-party content being incorporated into AI prompts and putting businesses on the wrong side of the law.

As a result, the CLA said it has expanded its corporate and public sector licences to include new workplace generative AI permissions, due to take effect from 1st May.

These permissions will allow licensed organisations to use published content in prompts for approved enterprise versions of genAI tools and hopefully ensure creators and rightsholders are fairly remunerated while still allowing businesses to leverage the benefits of the technology.

Recommended reading

• Action Needed Around Copyright and Competition in AI, Open Letter Says
• Class Action Lawsuit Could Steer AI Art Copyright Debate
• AI Could Improve Copyright Arbitration, New Study Says

“Generative AI, like previous technological advancements, presents significant benefits for our customers but also introduces new risks and challenges regarding the lawful use of copyrighted material by employees,” said Mat Pfleger, CEO of CLA.

“Our goal has always been to empower our customers to legally copy, use, and share copyrighted content while safeguarding the rights of creators and rightsholders.

“Starting in May, our corporate and public sector licences will aid organisations and their employees in leveraging GAI technology responsibly. This will help ensure fair compensation for rightsholders and support the continued flourishing of creative industries.”