Even as UK PM Rishi Sunak is pushing for the AI, his advisor on Artificial Intelligence (AI) Matt Clifford has raised concerns that AI technology – in future – could become powerful enough to cause significant harm to humanity within a short span of two years.
Clifford emphasised that even the immediate risks associated with AI were “pretty scary,” as the technology has the potential to create cyber and biological weapons.
The worrying revelation comes as Rishi Sunak, the UK Chancellor of the Exchequer, prepares to visit US in order to persuade President Joe Biden of his “grand plan” for the UK to play a prominent role in worldwide AI regulation.
One of the UK Government’s primary goals is to establish an AI Watchdog on the lines of International Atomic Energy Agency. In addition, the UK also intends to propose the formation of a new international research body. Clifford believes that without worldwide regulation of AI producers, the creation of “very powerful” AI systems would be difficult to manage.
This warning from Clifford follows a letter endorsed by numerous AI experts, including pioneers in the field, which stressed the urgent need to address the risks associated with AI on par with pandemics or nuclear war.
Prominent figures such as senior executives from Google DeepMind, Anthropic, and Geoffrey Hinton, often referred to as the “Godfather of AI,” have all expressed concerns about the potential dangers of AI falling into the wrong hands. Geoffrey Hinton recently resigned from his position at Google, citing AI’s potential to bring about the end of humanity.
Clifford, as an adviser to Sunak, is participating in the formation of a government task force to investigate AI language models such as ChatGPT and Google Bard. In addition, he is the chairman of the Advanced Research and Invention Agency (Aria). Clifford emphasised his fears about the current threats linked with AI in an interview with ‘TalkTV’, suggesting that the technology might be used today to manufacture bioweapons or conduct large-scale cyber attacks. He also stressed that this could be disturbing and harmful to the society.
Moreover, Clifford observed the concept of existential risk; similar to what the experts were referring to, stating that it pertains to the creation of a new species or of intelligence that could surpasses human capabilities.
While recognising that the idea of computers surpassing human intellect in two years is on the “bullish end of the spectrum,” Clifford pointed out that AI systems are always improving at a rapid pace.
He said, “I do not think it is zero.” He went on to say that the predicted advancement of AI models over the next two years could pose significant threats to humans in the form of bioweapons or cyber attacks, potentially resulting in the loss of many lives. He further concluded that the focus should be on developing mechanisms to control AI models effectively, as the current understanding of how to manage these systems is still narrow.
Meanwhile, Sunak’s father-in-law and Infosys Chairman N. R. Narayana Murthy had recently observed that with the introduction of AI, our lives could be made lot easier but he brushed aside the threats that it may replace humans. Murthy strongly believed that humans have the power of the mind and no computer can compete with that. Perhaps, Sunak could take a leaf out of his father-in-law’s suggestion.