New Research Debunks AI Doomsday Myths: LLMs Are Controllable and Safe

Large language models like ChatGPT are unable to learn or develop new abilities on their own, so they do not pose an existential threat to humanity. According to recent research from the University of Bath and the Technical University of Darmstadt in Germany, ChatGPT and other large language models (LLMs) are unable to learn autonomously [...]

from SciTechDaily https://ift.tt/okZIX9p