The UK’s national standards body, the British Standard Institution (BSI), is reportedly set to unveil a new standard for companies that independently audit artificial intelligence tools in a bid to protect innovation.
The new standard, tagged the UK AI audit standard, will focus on addressing unregulated and potentially risky practices in the emerging AI sector by introducing a structured framework for auditing AI systems.
The standard comes into effect on July 31
According to the BSI, there are now “hundreds” of “unchecked” groups offering audits that claim to assess whether companies utilizing AI models used in both traditional and pioneering use cases do so reliably, fairly and safely.
Many of the groups that sell AI audits also develop their own AI technologies, “raising concerns about independence and rigor”, the BSI said.
According to the institute, the standard will launch on July 31 and is the first international set of requirements to standardize how assurance firms verify if companies are complying with AI management standards.
The news comes weeks after the UK’s Financial Reporting Council (FRC) published its guidance on AI in audit on June 26, 2025. The guidance provided a coherent approach for implementing AI tools in audits, and also emphasized documentation requirements to support innovation in the audit profession.
The UK’s new standard is crucial
While the world has embraced AI, its vulnerability to mistakes such as hallucinations, and other dangers the technology poses has raised the stakes on reliable AI assurance services. Companies also need to worry about complying with international regulations, such as the EU AI Act, making assurance services crucial.
Boutique companies have come out of the woodwork to take advantage of the increase in demand, pitting themselves against larger contenders, including the Big Four accountancy firms.
Although it’s still a fledgling sector, the AI assurance market already generates a gross value of more than £1 billion for the UK. So, it makes sense that regulators have sounded the alarm on the lack of standardization in the sector.
That means a company can still offer assurance services that ends up being some light-touch advice, or just limited to checking whether the AI is compliant with one particular piece of legislation.
Mark Thirlwell, global digital director at the BSI, said: “Businesses need to be sure that when their AI management system is being assessed, it is being done in a robust, coherent and consistent manner.”
It is the BSI’s hope that the standard will aid regulators, customers and investors by helping them tell between AI that has been assured by a certified assurance provider, thereby “supporting responsible AI innovation.”
The standard is now being branded as a “pivotal step forward for the AI assurance ecosystem” as it clarifies which companies are qualified to certify AI systems against ISO standards.
Inioluwa Deborah Raji, researcher at UC Berkeley who specializes in AI audits and evaluations, has highlighted that “many assurance firms used proprietary systems to audit AI” and while companies “will pay to evaluate against [proprietary] standards”, she warned that there is currently no way of externally vetting the quality of these standards.
KEY Difference Wire: the secret tool crypto projects use to get guaranteed media coverage
