Digital technology companies have reshaped our world and will continue doing so. Sir Keir Starmer knows his government must seek a role in shaping this new order – and avoid ceding all control to the US and China. According to official estimates, the UK is the third-largest AI market. Its universities are important incubators of talent. Google DeepMind, two of whose scientists won a Nobel chemistry prize last year, was a British company until Google bought it in 2014. But the world’s two largest economies, and particularly the US corporations that dominate our online lives, are a long way ahead. The danger for the rest of the world is being swept along in an AI wave over which it has little control.
Expanding Britain’s publicly owned computing resource – a national asset known as sovereign compute capacity – is a necessary step toward securing technological independence. Wes Streeting, the health secretary, is right to warn of a coming battle to ensure democratic control of computing. Countering private-sector dominance with new public investment is part of that. Plans to boost the AI industry, beginning with a new growth zone in Oxfordshire, make sense as part of a 21st-century industrial strategy.
However, the enormous energy demands of AI data centres pose a serious challenge that cannot be ignored. In fairness, ministers know this, which is why they have announced a new AI energy council. But the energy secretary, Ed Miliband, and his environmentally minded colleagues, will have to fight hard for the energy impact of the policy to be taken as seriously as it should be by a prime minister and a chancellor whose overriding priorities are growth and increased productivity.
The government’s ambitions for AI-driven public sector reform raise another set of questions – although the risk of harms being ignored in favour of a boosterish focus on benefits is the same. Fragmented public procurement and a history of failed tech upgrades – including the Post Office Horizon scandal and a £10bn failed NHS patient record system – undermine confidence in the government’s ability to deliver meaningful AI reforms. Some experts believe the existing landscape needs to be tidied up before ambitious new schemes are launched. For example, there are key decisions about how much leeway organisations such as NHS trusts, councils or academy chains ought to have.
The “scan, pilot, scale” approach recommended by the tech investor who wrote the document on which the policy is based, Matt Clifford, skates over such challenges. While it acknowledges concerns over privacy, there’s no clear strategy for managing conflicts over sharing public data with private companies. And while ministers have not yet acceded to the industry’s demand for free rein to train machines on existing material (music, books and so on), it is disturbing that they have not ruled against this either. Having their work given away to internet businesses is the last thing that the UK’s embattled creative industries need.
There is no doubt that AI can do good. The hope of ministers that it can enhance public services, for example by giving GPs more time to talk to patients while computers take notes, is both sincere and realistic. But the ongoing difficulties faced by governments all over the world in regulating social media are a cautionary tale. Big tech talks a good game with the help of a formidable lobbying operation. Governments must cut through industry noise and prioritise responsible AI governance.
Source link
lol
