The UK offers a soft middle ground between Europe’s onerous approach and the laissez-faire US, where most AI innovations come from. The UK has been moving fast on technology governance before. The UK’s Children’s Data Protection Regulation is the first in the world, and was copied by California. The online safety law will come into force ahead of a similar set of rules by the EU. After Brexit, the company’s antitrust watchdog pursued a series of high-profile lawsuits against Alphabet’s Google, Apple and Metaplatforms.
Meanwhile, the UK is upgrading its government machinery to tackle AI more effectively. For example, the UK recently consolidated its top regulators for coordinating decision-making into a single body. This will be essential for technologies impacting nearly every industry. In addition to a strong commercial law field and a centuries-old reputation for the rule of law, the UK has one of the world’s largest AI companies, Google DeepMind, where technologists work with governments as advisors. We also have culture. No problem if everyone speaks English.
“There is only one positive aspect of Brexit, and that is that the UK is in a position to become a world leader in AI governance,” said Saul Klein, co-founder of London venture capital firm LocalGlobe. In her February 2023, the UK will combine the government departments responsible for business and digital sectors to be called the Ministry of Science, Innovation and Technology. Klein, the non-executive director of the new division, said technology has become as important to the government as the economy and national security.
The UK has also brought together the best regulators dealing with issues such as antitrust, data privacy and online harm under one umbrella called the Digital Regulatory Cooperation Forum. This makes it much easier to set and enforce rules around AI with one voice. Technology regulation expert Rachel Cordicut says why spend years building a new regulator when existing oversight agencies are rising together to tackle the tech disruption . He specializes in technology regulation and is the Executive Director of Promising Trouble, a social enterprise that builds and supports alternatives to Big Tech. .
How will the UK seek to regulate AI? Upcoming rules on online hazards borrow principles from the Health and Safety Act, imposing a legal “duty of care” on businesses to maintain a safe environment. giving to The UK will likely follow a similar path with AI, adopting a risk-based approach to regulation, according to regulatory experts.
DeepMind co-founder and government adviser Demis Hassabis shares a similar view. He recently advocated a “precautionary principle” in regulating AI. It essentially means taking precautions when you don’t know the risks. For example, like telling a child not to play soccer in the house so as not to break something. An alternative is to let the kids play and make new rules about soccer in the house only if they break the lamp.
But there is one glaring problem with Britain’s efforts to oversee AI, and it speaks to the country’s tendency to be friendly and insular. Sunak recently announced a global AI safety summit this fall to kick-start his country’s regulatory ambitions, but he ignored the participation of those working in civil society.
According to the 1,400-word press release announcing the event, the summit will bring together countries, researchers and “big tech companies” such as Google DeepMind and Microsoft Corp. But nowhere in the celebrities’ names was there any mention of people studying flaws in current AI systems. A British government spokesman said the summit’s aim was to build “international guardrails for the safe and responsible development of AI”.
The current AI problem is serious and pervasive. From algorithms that check asylum seekers’ applications to algorithms that filter job ads, they frequently affect women, people of color, and other minority groups. UK organizations such as the Ada Lovelace Institute, Demos, Connected by Data and Promising Trouble, which investigate such flaws, should also be invited to discuss AI safety with major companies.
“Our experience over the past 20 years has shown us that asking technology companies to self-regulate is ineffective, and the idea that technology companies are safety experts is ludicrous,” Cordicut said. increase. “What we really want at the AI Safety Summit is for people to talk about the world we want.”
The UK already has the infrastructure and reputation to take a decisive approach to AI governance. But Sunak must resist the attention of big players like OpenAI and make sure his policies take into account the problems on the ground. He should listen to experts on those issues as well.
Details from Bloomberg Opinion:
• At Apple, we trust more and more personal data: Palmy Olson
• Reddit made the mistake of ignoring core users: Dave Lee
• Sam Altman asked for AI rules. EU Answer: Lionel Laurent
This column does not necessarily reflect the opinions of the editorial board or Bloomberg LP and its owners.
Palmy Olson is a technology columnist for Bloomberg Opinion. She is a former Wall Street Journal and Forbes reporter and author of We Are Anonymous.
More articles like this can be found at bloomberg.com/opinion.
