How does artificial intelligence avoid unclear answers?

🔹Google DeepMind researchers showed in a four-stage experiment that large language models can assess their level of confidence in answers and refuse to provide answers if they are not confident enough; A feature that helps reduce misleading answers and increase the reliability of artificial intelligence.

🔹 The TechExplorer website said in the report:

Big Language Models (Alam) are the same AI systems behind platforms like Chatjipit and Gemina. These models are expected to answer questions, generate text and make decisions without the supervision of human experts; But previous studies show that these models sometimes mislead users and provide uncertain answers; Answers in which a lot of confidence is expressed in false or inaccurate statements.