Tech entrepreneur Matt Clifford is urging immediate action to regulate the capabilities of AI, citing concerns that failure to do so could result in the development of bio-weapons with the potential to cause significant loss of lif
Last week, a letter signed by 350 AI experts emphasized the need to treat AI as an existential threat to society, akin to pandemics or nuclear weapons.
During an interview on Talk TV, Clifford explained that the signatories of the letter are expressing the view that AI systems are advancing at an exponential rate and becoming increasingly capable. Without proactive regulation and safety considerations, he warns that within two years, we could be facing highly powerful systems.
When discussing the timeframe for addressing these issues, Clifford acknowledged the wide range of predictions among AI experts. He believes that two years represents the most optimistic end of the spectrum.
Prominent figures from companies such as Google, DeepMind, and OpenAI have joined Clifford in calling for the risks associated with AI to be treated as a global priority due to the potential threat they pose to humanity. Elon Musk, CEO of Tesla and SpaceX, has also voiced concerns about the potential harm caused by "Terminator" style robots.
Clifford, who chairs the government's Advanced Research and Invention Agency, described the world as being at a tipping point, with even short-term risks being deeply concerning.
Speaking on the First Edition show, Clifford stated that the creation of AI that surpasses human intelligence without adequate control mechanisms presents various risks, both presently and in the future. He believes that policymakers should prioritize addressing these risks.
While acknowledging that the risk of AI turning against humans may sound like the plot of a movie, Clifford emphasized that it is not zero.
He highlighted the striking rise in AI capabilities and the potential dangers posed by bio-weapons and cyber threats, which could result in the loss of many lives within the expected advancements of AI over the next two years.
Furthermore, Clifford mentioned that the signatories of the letter openly admit their lack of understanding regarding the behaviors exhibited by machines, which he finds quite terrifying.
