The AI hype cycle may be hitting its first major slowdown as new data reveals growing skepticism among business leaders. Concerns over implementation risks and unclear returns are tempering initial enthusiasm for rapid adoption.

Fresh insights from global insurance firm Gallagher’s second annual Attitudes to AI Adoption and Risk Survey highlight a growing wariness around AI adoption, with 11% of business leaders now viewing AI as a potential risk, up from just 5% a year ago.

However, Gallagher’s survey of 900 business leaders worldwide found that AI risk awareness remains steady, with 77% understanding the risks, slightly down from 78% last year, although the communication of these risks to employees has declined to 78%, down from 84%.

AI hallucinations – where systems fabricate facts or information — tops the list of risks, with around a third (34%) concerned about the impact that inaccurate AI outputs might hold. 

A third (33%) of business leaders also reported being worried about the increased threat to privacy and the possibility of data breaches that increased AI use might bring, while 31% said they were concerned about the legal liability of AI misuse, and 29% were fearful of the vulnerabilities to cyber-attack and fraud more AI might cause.

Only 9% of those questioned said they could not see any potential risks connected with using AI.

According to Gallagher’s polling, to address these challenges, many businesses have begun to update their risk management frameworks to include AI-specific challenges, with nearly half having revamped risk strategies to include AI considerations.

Rising fears over cybercriminals’ ability to leverage AI are behind the 42% of respondents who said they’ve strengthened their cybersecurity practices, while 41% have reassessed their privacy and data security measures to mitigate AI risks.

Gallagher said it predicts there will be an increase in legal liabilities for firms where the use of AI is not properly governed. 

Recommended reading

• GenAI Rollout is ‘Tearing Companies Apart,’ New Report Claims
• AI at Work Sparks Copyright Concerns for UK Firms
• What Key GenAI Advancements Will Shape the Future of Procurement?

One particular area of concern is firms or contractors relying on AI based research to provide professional services or advice to clients. If they use information which turns out to be incorrect, these businesses are at significant risk of having to pay out legal costs, expenses, damages, and settlements to clients.

Firms should also be aware of the legal exposures that AI might bring to privacy and data protection risks, with businesses misusing AI open to individual actions and potential class actions for privacy violations,  including data usage rights infringement, data poisoning and regulatory infringement. 

“Historically, the cyber insurance market has been adept at adjusting to new threats over time, and the expectation is that it will do the same with risks related to AI,” said James Pearce, cyber account executive at Gallagher.

“However, as AI risks become more defined, insurers are starting to review their cover on policies, and businesses using AI need to check if they do have coverage if a cyber-attack or privacy violation due to AI exposures were to take place.”