One researcher’s experience coaxing multiple chatbots into explaining firebombs, poisons, and uranium enrichment—and then being ignored by vendors and even government agencies—underscores how unprepared our institutions are for real-world AI misuse. https://spectrum.ieee.org/jailbreaking-llms?utm_source=mastodon&utm_medium=social&utm_campaign=fedica-Mastodon-Daily-Pipeline