A normie’s guide to understanding the global AI safety debate
AI safety is becoming a geopolitical and public policy maze. A return to first principles may show us how to erect sensible guard rails.
Sign up now: Get ST's newsletters delivered to your inbox
Sensible guard rails can reassure ordinary citizens that AI will not be allowed to run unchecked, says the writer.
PHOTO: REUTERS
With a torrent of specialist jargon – from “recursive self-improvement” and “p(doom)” to “pacing the frontier” – the snowballing global debate over artificial intelligence safety is becoming harder to follow by the day.
If it is not these terms sending us normies – normal citizens outside the AI world – scrambling to keep up, it is the proliferating catalogue of doomsday scenarios. These range from mass unemployment and DIY cyberattacks or bio-weapons, all the way to the ultimate fear: human extinction.