There is some truth in the notion that our national psyche dwells more on problems than solutions. A tendency that arguably runs against entrepreneurialism, especially in the AI age when speed is king.
But there is another way of looking at it. Britain has a long history of turning risk into business.
A market waiting to be made
Every major economic shift has created markets around trust.
The Big Bang opened London to international competition and capital, increasing the need to give investors confidence in how those markets operated. Britain developed global strengths in law, insurance, assurance, and audit and financial regulation – capabilities that were further bolstered after the financial crisis. With the digital economy came cyber security, identity and payments. None eliminated risk; they made it more manageable. Advanced AI will create its own trust economy, and the UK has a huge opportunity to lead it.
We won't train out every risk
Britain is rightly investing in sovereign AI capability, but many of the most powerful language models will continue to be built elsewhere; their characteristics shaped long before a business gets its hands on them.
Even our finest AI minds cannot expect to train every unwanted behaviour out of such models once they are in the field. Nor can we assume that putting a sufficiently strong harness around these models will completely constrain what they do.
What we can do is test models under stress, identify their strengths and weaknesses, understand behavioural traits and establish where additional controls are needed. That’s why the AI Minister's recent confirmation that the UK will retain early access to the most powerful models for testing matters so much. We can then advise on the governance, monitoring and safeguards so users encounter less unwanted behaviour and fully understand the risks.
Giving boards confidence
That understanding could give boards greater confidence to use AI for more consequential activity, in turn creating greater value. PwC’s latest CEO Survey found 56% of CEOs have yet to see either revenue or cost benefits from AI. The barrier is often not the technology, but confidence in it. Boards are asking which models they can rely on, for which decisions, and how they'll know when something goes wrong. Answering those questions pays off. Separate PwC research shows that the leading 20% of companies are already capturing nearly three-quarters of AI's economic value. They are also 2.8 times more likely to be increasing the number of decisions made without human intervention – while being more likely to have cross-functional governance in place.
The UK has the right ingredients
Britain is unusually well equipped to do this well: we’re home to the AI Security Institute, leading universities and researchers, regulators and standards experts, alongside world-leading capabilities in law, accounting, insurance, cyber security, risk and assurance.
We can turn that collective instinct for understanding risk into something the world values – and buys. In doing so, we have an opportunity to act as a trusted global bridge, open to innovation wherever it is developed.
"The UK should not treat AI safety and trust as a defensive strategy sitting alongside its growth strategy. Trust can become a source of AI growth in its own right. Britain has built globally successful industries on trusted institutions, expertise and judgement – we should seize the moment to do this again.”
Claire Reid, Chief Technology and Innovation Officer, PwC UKThe prize is bigger than AI assurance alone
The Government estimates the AI assurance market could exceed £18.8bn by 2035 if adoption barriers are addressed. Assurance is not the end market. It is part of the infrastructure that allows a much larger AI economy to function with confidence.
The bigger opportunity is what trust can unlock across the economy, such as faster investment decisions, more valuable use cases, new products, and global firms choosing the UK to evaluate and deploy advanced technology.
There is a further prize.
Influencing the technologies people choose
If global businesses trust UK expertise on where different models perform well and where they don’t, as well as the circumstances in which they can be deployed safely, that judgement can start to influence which technologies they choose.
And if trusted judgement helps determine those choices, model developers have a powerful incentive to respond to what those customers value.
Creating the market
The next step isn’t another set of responsible AI principles. We need a functioning market for trusted AI, connecting Britain's safety science with the standards, evidence and independent assurance businesses need to make real decisions. We have the expertise and just need to turn the disparate parts into a system, and one the world wants to use.
The ultimate opportunity is for Britain to become as influential in judging AI as others are in building it.
Contact us
