Meta withdrew its artificial intelligence for science Galactica, which it introduced earlier this week, amid strong backlash. Galactica was introduced as an AI language model that summarizes research papers, solves equations, and performs a host of other useful scientific tasks. Scientists and academics have realized that the AI system generates a lot of misinformation, including citing real authors for non-existent research papers. Galactica was posting misleading or completely false scientific papers and was inconclusive on a variety of issues.Great threat to science!
“We have evaluated the feedback we have received from the community so far and have paused the demo for now.” made the statement. “The Galactica demo is closed for now,” tweeted Yann LeCun, Meta’s chief AI scientist.
Galactica is a major language model, a type of machine learning model known for writing highly convincing text that feels like it was written by humans. But artificial intelligence isn’t very good at understanding human language yet. This is a big problem when it comes to scientific research. Scientific papers rely on immutable methodologies that writing AI systems cannot clearly understand, or at least not yet grasp. Researchers see systems like Galactica as the beginning of a new era in which scientific misinformation will spread.
34733
