Some of the UK and Ireland’s largest organisations are looking to their software vendors to help them stay on the right side of AI regulations, research has revealed.

Two-thirds (66%) of respondents surveyed by SAS, a global AI and analytics firm, said they expect their current AI software suppliers to help them comply – placing more trust in them than their own legal (57%) and IT (21%) teams.

According to the survey, they are also more than twice as likely to trust their current vendor compared to a new one.

As many as 85% said they were confident that their AI software could meet new regulatory requirements – and the same proportion again were confident in their vendor’s knowledge of the pending legislation.

Over half (57%) said this was because their vendor had shared their expertise with legislators ahead of new laws being drafted, and 54% said their vendor had previously advised them on AI compliance.

The study also found that 88% of organisations had already started to make changes to their software in line with new legislation or had plans to do so.

That leaves 12% who have yet to take action, putting them at risk of non-compliance and falling behind the technology curve.

Recommended reading

• Exploitation of GenAI a Stark Concern for IT Experts
• Survey: UK Workers Embrace Generative AI But Lack Safety Skills
• Are UK Consumers Ready for Businesses to Adopt AI?

Lack of resources, such as skills, was the reason cited by 72% of respondents, while half (48%) believe they lack third-party support.

However, this trust may be misplaced: a new transparency index out of Stanford University found that almost all major AI companies fell below par, meaning even the vendors may struggle to keep up with new regulations.

The report comes as the EU is expected to implement its AI Act by March 2024, the first of its kind by a major regulator.

Further, the UK government has set out five principles for the responsible use of AI, stressing that it wants to avoid formal legislation. However, it has yet to rule out introducing a statutory duty for regulators to ensure the principles are followed.

“AI is arguably the most powerful and transformative technology we’ve seen in a generation – so it’s more critical than ever that organisations exercise due diligence when selecting a vendor. They need a trusted partner who can advise on adoption of best practices across the organisation, and has the ability to create trustworthy AI solutions,” Dr Iain Brown, head of data science at SAS Northern Europe, said.

“In addition, it’s important to foster accountability throughout the company. Staff training and consciousness regarding data usage need to be firmly grounded in data management principles, including data quality and lineage. It is also crucial to establish robust model governance to continually monitor the array of models being used across an organisation, and ensure outputs minimise potential bias.”