Rishi Sunak revealed plans for a UK-based institute focused on AI safety, but has rejected the idea of halting progress on the technology’s advancement.
The institute, according to the prime minister, will be a pioneering venture focused on experimenting with various forms of AI to address a wide range of risks, including spreading false information and potentially posing a major threat to humanity.
Sunak has revealed that the institute will be announced prior to the upcoming global summit on AI safety at Bletchley Park next week. He also stated that the purpose of the institute is to progress understanding and research on AI safety on a global scale.
In a speech at the Royal Society, an organization of prominent scientists, he stated that they will thoroughly analyze, assess, and experiment with different forms of AI in order to comprehend their capabilities.
According to him, the institute will thoroughly assess all potential risks associated with AI, including societal issues like prejudice and false information, as well as the most severe risks. The UK’s frontier AI taskforce, which was formed this year to examine the safety of advanced AI models, currently has a prototype of the safety institute in place.
Sunak stated that it would not be possible to halt the development of advanced models. When asked about his stance on implementing a temporary ban or restriction on the creation of artificial general intelligence, he expressed that it would not be feasible or enforceable. He believes that the UK should continue to foster innovation for the benefit it brings, as it has historically done.
In March, the discussion surrounding the safety of artificial intelligence reached a significant milestone with the release of an open letter signed by numerous influential figures in the technology industry, such as Elon Musk. The letter urged for a temporary halt in the development of “giant” AIs for a minimum of six months. It expressed concerns that AI labs were in a frenzied competition to create and implement increasingly advanced digital minds that even their creators could not fully comprehend, predict, or effectively manage.
Sunak stated that it remained uncertain if China would participate in the summit, despite being invited along with technology executives and other international leaders. The prime minister expressed that he could not guarantee with absolute certainty that Chinese officials would be in attendance.
In May, a statement signed by leaders in the tech industry was referenced by Sunak on Thursday. The statement discussed the importance of addressing the potential dangers of AI development, comparing them to the threats posed by nuclear weapons and pandemics.
Some AI experts are concerned about the possibility of AGI, which refers to a system that can perform a variety of tasks at or above the level of human intelligence.
Sunak addressed the public following the government’s publication of its evaluation on the safety hazards of AI, which included acknowledging the possibility of an existential danger posed by the technology.
According to a government document released on Wednesday, due to the uncertainty surrounding the development of AI, there is not enough evidence to dismiss the possibility that advanced AI systems could present a threat to humanity if they are not properly aligned or controlled.
Other dangers outlined in the risk documents from the government included the capacity for systems to create bioweapons, produce large quantities of “hyper-targeted” false information, and significantly impact the job market.
Sunak stated that the possibility of a “superintelligent” system posing an existential threat and operating beyond human control is a controversial topic among experts. However, he acknowledged that developers of advanced AI have expressed concerns about potential existential risks.
“Even though these risks may be uncertain and improbable, the potential consequences would be extremely severe. Given that even major technology developers are acknowledging these risks, it is crucial for leaders to acknowledge and address them seriously.”
Sunak stated that he plans to utilize the summit on November 1 and 2 to advocate for the creation of a professional AI observation team modeled after the Intergovernmental Panel on Climate Change.
He stated that he plans to suggest the creation of an international panel of experts, chosen by the countries and organizations attending the summit, to release a report on the current state of AI science next week.