AI Flags Threats From Human Extinction to Bioweapons

In its September 18 newsletter 'The Download,' MIT Technology Review examined the extinction risk posed by AI and the potential threat of bioweapons (weapons that use pathogens, toxins, and the like). Could AI wipe out humanity? This issue features the questions and answers from Wednesday's live Roundtables event, 'Can AI really kill us?' Senior AI editor Will Douglas Heaven and AI reporter Grace Huckins fielded questions such as 'Is AI dangerous, or is this just a publicity idea on the part of tech companies?' and 'What steps are needed to effectively control, monitor, and regulate AI?'

The 2022 study was cited as evidence of the bioweapons threat. Using a drug development AI known as a 'molecular generator,' researchers produced 40,000 molecules that could be used as chemical warfare agents within six hours. The newsletter explained that the risk is growing as AI tools are now able to answer questions in nearly every field of science and gene-editing technology continues to advance.

This issue also included remarks by Microsoft and OpenAI employees saying that 'AI is destroying the web,' as discussion continues over the various risks posed by AI.