
South Korean researchers develop technique that enables artificial intelligence to refuse providing false information when it does not know the answer
ℹ️ Browser-based reading · AI studio voice coming soon
Researchers from South Korea have developed a new method that enables artificial intelligence systems to acknowledge when they do not know something rather than producing misleading information.
The development comes amid efforts to improve the safety of artificial intelligence applications, particularly in sensitive fields such as healthcare and autonomous driving. The new technique aims to reduce errors in large language models.