Despite the transformative potential of Generative AI (Gen AI), business leaders are increasingly concerned about its impact on business performance, according to new research from KPMG UK. The study reveals that while there is growing awareness of the risks associated with Gen AI, efforts to address these concerns through clear guidelines and education are falling short.

Some 60% of respondents identified the inaccuracy of Gen AI results, including the phenomenon known as "hallucinations"—where AI generates false or misleading information—as their top concern. Other significant worries include errors in the underlying data that could distort AI outputs (53%) and cybersecurity risks (50%). These concerns reflect a broader unease about the reliability and safety of deploying Gen AI in business operations.

But only 30% of directors reported that their organizations have established and communicated responsible Gen AI usage guidelines to mitigate these risks. This slow response is particularly alarming given that, in a separate KPMG poll, 42% of UK adults admitted to inputting work-related information into Gen AI tools, underscoring the urgent need for clearer guidance in the workplace.

Algorithmic bias is another pressing issue, with 43% of board members expressing concern about this potential pitfall. Yet, a strikingly low 8% of organizations have implemented processes to measure and address bias in their AI systems, leaving a significant gap in ensuring equitable and accurate AI outputs.

Leanne Allen, Head of AI at KPMG UK, said: "Given boards’ concerns, it’s important that companies thoughtfully define a clear AI strategy rather than merely chase the next technological innovation. This strategy should balance the value, cost, and risk associated with AI use cases. This strategic equilibrium is crucial for both progress and stakeholder trust," she said.

On a more positive note, nearly a quarter (23%) of businesses have begun implementing or are developing Gen AI training programs to address these risks. However, most board members are still learning about Gen AI through informal means, with 68% relying on ad hoc or self-taught methods, and 48% depending on management presentations or third-party training. Only a small fraction of boards (5%) have, or are actively seeking, Gen AI expertise.

UK boards appear to be slightly ahead of their US counterparts in recognizing the need for Gen AI expertise, with 18% reporting they had such expertise compared to just 10% in the US.

As Gen AI continues to evolve, the challenge for businesses will be to not only harness its potential but also to ensure that robust guidelines, education, and expertise are in place to navigate the associated risks responsibly, said KPMG.