MIT Technology Review’s daily technology newsletter discusses the possibility that advanced artificial intelligence could wipe out humanity. The subscriber-only Roundtables meeting, featuring Niall Firth, Will Douglas Heaven and Grace Huckins, will be held Tuesday, September 15, at 16:00 BST / 11:00am EST / 8:00am PST. The discussion will address the sources of the risks and possible precautions.
A “reprogramming” method developed by geneticist Yuancheng (Ryan) Lu restored sight in blind mice by repairing their optic nerves; a similar genetic treatment has entered human clinical trials. Known for his research on aging and the eyes, Lu was included in the 35 Innovators Under 35 list for 2026.
Dario Amodei, Sam Altman and Elon Musk called for slowing the development of artificial intelligence and imposing stricter oversight; while Amodei proposed independent monitoring and Altman advocated controlled progress, Musk said, “Dario is right.” Following these calls, AI-related stocks declined, while Chinese state media described the statements as a “Cold War” tactic. Trump and the Speaker of the House opposed stronger regulations, while Democrats want new rules. Xi Jinping proposed open-source AI cooperation among BRICS countries. South Korea increased penalties for espionage to 30 years to protect chip secrets. The US and Mexico plan to use high-energy lasers against drones at the border. The independent actions of AI agents, meanwhile, are creating legal uncertainty for the criminal justice system.
Why it matters
The content shows that the debate over artificial intelligence is not limited to the capabilities of models, but extends to areas such as safety oversight, government policies, chip information and border security. For this reason, the issue concerns not only technology companies and researchers, but also regulators, legal experts and public institutions affected by artificial intelligence systems. The transfer of results from blind mice in genetic treatment to human clinical trials indicates that the research has moved from a laboratory finding to clinical evaluation; however, the text presents no conclusions regarding efficacy and safety in humans. The open questions are how different countries and political actors will establish joint oversight for artificial intelligence and what results this genetic approach will produce in humans.