The UK’s much-hyped AI Safety Summit is upon us. At a time when global events could hardly feel more urgent, world leaders and industry leaders are looking to the future. But it is a future that could be even more hopeless than the present day.
Will we use AI to make new scientific discoveries? Explore the edges of the universe? Find cures for cancer? Maybe. But the government is using the Summit to rally the world to dwell, somewhat aimlessly, on the risks of an AI apocalypse.
The UK government aiming to be the AI safety leader of the world may sound like a platitude to those of us raising the alarm on AI risks, but it misses the mark by a long way. The risks of frontier AI being misused or escaping human control should be considered, but the most urgent risks posed by the technological revolution are being overlooked. Major technological changes are already underway – and are changing our country.
Take the UK’s farcical experiment with AI live facial recognition surveillance. In London and Wales, 89% of live facial recognition flags have misidentified innocent people as those on police databases.
And yet this week, the Policing Minister called on police forces to “double” their use of AI facial recognition surveillance. The technology will improve – but I’m not sure that the abject failure is worse than the prospects of this technology working perfectly. Do we want our streets to become perpetual AI police line ups? Should technology permit governments to treat citizens like suspects?
Chris Philp wants the Orwellian AI tech to be fitted to police body-worn cameras and public CCTV cameras to “check the identities of people of interest encountered on the streets in near real time”. He recently pledged to turn every single passport photo into a police mugshot, enrolling 45 million of our photos in police databases alongside criminals. It’s almost like this nation’s revolt against national ID cards – opposition to which was one of the Conservative’s flagship policies helping them win power in 2010 – has been consigned to the memory hole. Is Britain’s picture of the future AI cameras scanning every human face, forever?
Despite posing real and current threats, AI surveillance seems to be forgotten in the ‘AI safety’ agenda.
The result is this. In a country that proudly enshrines the presumption of innocence in law, at least twice as many people have been wrongly stopped by law enforcement as a result of AI facial recognition uses over recent years than those correctly stopped. Not a single law in the UK contains the words ‘facial recognition’ and the technology, with its numerous safety risks, has never once been debated in the ‘mother of parliaments’. Whilst taxpayers are footing the multi-million pound bill for this extraordinary technology that changes the relationship between the citizen and the state, it has evaded every possible democratic process. And holding an AI Summit to swap notes with China about it does not count.
The perilous international landscape should make us see the flaws in the UK’s current outlook on AI more sharply than ever. The world is divided into new political axes – the authoritarian axis of Iran, Russia, China and North Korea is looming over the liberal West. Its collective strength, both militarily and ideologically, poses an existential threat to Western liberty.
In a world increasingly divided between authoritarianism and liberalism, adopting authoritarian technologies comes with a real cost. The British government cannot line the streets with AI CCTV-watchmen and facial recognition software; or cover our schools and hospitals with Chinese Hikvision AI cameras infamously used in China’s concentration camps; or monitor millions of ordinary people’s social media posts with an AI Big Brother to detect and censor ‘mistruths’ – all of which they do – and remain a liberal democracy with a population that has the strength to face down a growing axis of authoritarian evil. Nor can we preach to the darker sides of the world about how to embrace technology to build a freer future.
In these turbulent times, we all need to know what we are living for – and the values that we may need to fight for. The AI summit is about safety, but we cannot be complacent about liberty – there is no safety in a world without liberty.
— Silkie Carlo
Director, Big Brother Watch