The AI safety summit organized by Rishi Sunak may seem well-organized, but upon closer inspection, warning signs arise. Written by Chris Stokel-Walker.
The UK’s AI safety summit begins this week at Bletchley Park, organized by Rishi Sunak, the Prime Minister who is seeking positive news amidst the looming threat of a major election loss.
Sunak aims to leave behind a lasting impact in the field of AI. In a recent address, he discussed the potential hazards of AI being used by terrorists and cybercriminals. He also released a set of materials on “frontier AI”, which refers to generative AI technology like ChatGPT and DALL-E. Additionally, he introduced a UK institute dedicated to AI safety.
The statement was unambiguous. The smooth, although significantly trailing in the polls, Stanford MBA graduate who enjoys vacationing in California had, as he often says, “taken control” of the issue. Sunak stated that the British citizens can be assured that we are creating the most cutting-edge safeguards for AI compared to any other nation in the world.
The summit is next on the agenda. Sunak has managed to gather 100 prominent figures in the field of AI at Bletchley Park, including delegates from major tech companies and leaders from various countries. However, there is a significant drawback – there is a lack of representation from civil society, which has caused frustration among over 100 individuals who signed an open letter released on Monday. They have expressed concern that the meeting will not be successful with such a limited list of attendees.
It’s difficult to contradict their perspective. The summit’s main focus is on the potential dangers of an AI surpassing human intelligence, similar to the scenario depicted in Terminator. However, numerous experts and advocates in this field, who have been researching it long before it became Sunak’s pet project, believe that this risk is exaggerated and used as a diversion from the larger concerns surrounding AI.
According to warnings, there are other urgent issues to address, such as prejudice and discrimination against marginalized groups. A prime example is the lack of diversity in generated images of professionals like doctors and CEOs, which tend to only feature middle-aged white men. Additionally, the government’s plans to incorporate AI surveillance into police operations could further perpetuate these problems of underrepresentation.
The absence of AI’s environmental consequences on the event’s schedule is concerning, as it heavily depletes our planet’s resources. The amount of energy consumed by AI is projected to surpass that of numerous nations in a short period of time, yet it is only briefly mentioned in the discussion paper.
The language used in the promotional materials for the event, mentioning “frontier AI,” seems to heavily favor the current state of affairs. It mirrors the name of the Frontier Model Forum, a forum for the technology industry, hastily put together to appear self-regulating and prevent government regulation. The use of industry terminology suggests that Sunak will continue to be passive in the face of big tech’s latest advancements. This is not surprising, as the prime minister has openly expressed his desire to incentivize tech companies to develop AI in the UK in order to benefit the economy. He has also implied that he will not pressure them too much on safety regulations. In fact, Sunak is so eager to align himself with tech representatives that he will be joining Musk on X, formerly Twitter, for a livestreamed event after the conference.
In the past year, I have interviewed various professionals for a book about the significant influence that the current wave of artificial intelligence will have on our daily lives. This book is scheduled to be released next year. In the recent month, I observed government officials from the UK proudly discussing their country’s key role in regulating this technology in the future. During a conference in Amsterdam, I listened as Viscount Camrose, the UK’s Minister for AI and Intellectual Property, referred to the summit as “historic”. Additionally, I tuned into a live broadcast of Deputy Prime Minister Oliver Dowden building excitement for the event.
However, I have observed the individuals representing the tech companies and the government at the negotiating table. I have also noticed the sharp intellect of those who have not been invited to the summit, left in the cold this November. I am not optimistic about achieving favorable outcomes and a new agreement on artificial intelligence that effectively addresses our present challenges.
Chris Stokel-Walker is the author of How AI Ate the World, to be published in May 2024