AI systems develop deceptive capabilities

September 22, 2023 / Thilo Hagendorff

AI Insider: Every month, AI expert Thilo Hagendorff answers the most pressing questions about current developments in artificial intelligence.

AI Insider - September 2023

"It is one of the most important questions in AI safety: do language models have the ability to deceive humans? If so, this would pose major risks."
Article by Thilo Hagendorff published on September 22, 2023, 12:00 a.m.

Link to full article


Dr. Thilo Hagendorff

To the top of the page