The UK government's AI plan covers all the bases but needs a dose of pragmatism
With the launch of its AI Opportunities Action Plan, few people can complain that the UK government is not taking the potential of artificial intelligence seriously.
With 50 recommendations in the report by AI advisor and tech startup guru Matt Clifford, few can complain that the plan is not comprehensive.
With pretty much all of the obvious challenges included – skills, data sharing, investment, safety, regulation, datacentre capacity, energy availability, copyright and more – few can complain that the plan does not cover all the bases.
With industry leaders across the country offering warmly supportive statements and commitments, few can complain that the plan does not have widespread backing.
With public money behind the development of national infrastructure and “AI growth zones”, few can complain that the government is not putting our money where its mouth is.
And yet…
When I’ve seen criticism of the AI plan since its launch, my instinct as a technology advocate is to respond with optimism and to counter the detractors.
But when I’ve seen the boosterism around it – largely from people who stand to make a lot of money from selling AI – my instinct is to respond with caution, pragmatism, and the bitter realities of watching government once more turning to the Magic Technology Tree for solutions.
This time, perhaps the government hopes the difference comes in asking a robot to give the tree a shake. And I don’t mean Keir Starmer.
Yes, there is no doubt that AI is going to be the latest in a line of transformative technologies, with great potential to contribute to economic growth. Previous governments missed an opportunity with another transformative technology – cloud computing – where 15 to 20 years ago the UK could have become the datacentre capital of Europe, but failed to act.
In the coming years, AI will offer capabilities beyond what technology has previously been able to deliver – not least where huge scale and enormous complexity will always prove to be beyond the scope of us puny humans.
But if there’s one thing I’d really like to see, it’s for a one-year ban on anyone with a vested financial interest in selling AI from talking about it through the government. Leave the space open for people who have to buy, implement, deliver and use AI in real life to talk about its challenges and benefits.
If, in a year’s time, the public space is full of practitioners with experience of using AI to transform their organisations, their employees’ and customers’ lives, then let’s judge whether the government’s plans will really be what Starmer has called “the ultimate force for change and national renewal”.
AI is going to be great. But it’s still just a technology. It still needs people. It still requires organisational, cultural, social and personal change for any business, government department or individual seeking to exploit its potential. And that change is going to be hard. AI will follow the same, natural cycle of absorption into our lives as every other transformative technology that has come before. It will happen slowly at first and then, eventually, it will happen quickly. It will take time.
Like any technology, if you force AI onto an unwilling recipient, it will fail. If the loudest proponents of AI’s supposedly startling capabilities are correct, then where AI fails, it will fail with startlingly negative outcomes. Fears of an AI-enhanced equivalent of the Post Office scandal are justified.
AI is not a panacea. No technology has ever been a panacea. No technology ever will be a panacea.
Credit is due to Matt Clifford because in the comprehensiveness of his 50 recommendations, he has anticipated the most likely scepticisms of the AI naysayers, of which there are many. I suspect he has been there, seen it and done it many times when it comes to government technology plans that sound great in a press release but never quite deliver on their promise.
He’s a brave man in stepping forward to be the latest person to attempt to turn that around. He’s not the first – I doubt he will be the last. Whoever finally solves the conundrum of shaking the Magic Technology Tree and something falling out that is actually what they promised, deserves every reward that would come their way, because the rewards for government and public would be significant.
A dose of AI pragmatism and tech reality should be the 51st recommendation of the government’s AI plan, because Starmer is certainly right in one important aspect – this needs to be a government technology plan that succeeds.