Source: The Guardian

AI risk is described by Richard Hermer as unprecedented and potentially existential for humanity. Hermer says he spent two years as attorney general focusing on national security, but considers AI to pose a different magnitude of threat. He argues that the issue has recently gained traction after the disclosure of the Hugging Face hack and Jacob Coxon’s resignation from Anthropic. Hermer calls for greater urgency in international coordination to regulate frontier AI. He also says there must be meaningful public engagement on how AI should be shaped to benefit humanity rather than entrench the power of a few. Hermer frames the risks in two broad categories: existential risk and existential challenge. In his view, AI could enable the design and deployment of biological weapons or even decide it is more efficient to dispense with humans. He also warns that unchecked AI could deprive billions of meaningful agency over their lives, with major consequences for economies, national security, and human happiness.

Hermer says the rapid development of AI creates systems humans do not fully understand and may eventually be unable to control. He compares the situation to the 2008 financial crash, but says the current challenge is on a different scale. Hermer argues that frontier AI should be slowed down until risks can be managed, and rejects self-regulation by the industry. He says there is more regulation for a fridge than for frontier AI labs that could “wipe out your family,” and that this imbalance is startlingly wrong. Hermer says international law will matter most, and that the scale of the challenge requires international cooperation. He calls for engagement with the US, China, like-minded states, and the tech sector. He concludes that democracy requires urgent public debate so that AI serves the public good rather than the tech sector.