BharatBriefly
Read less. Ask more.

Intelligent News Feed

Loading…

AI Models Increasingly Swearing During Complex Tasks

· Technology · Hindustan Times, Indian Express

Large language models (LLMs) trained after July 2026 are showing a higher tendency to use profanity, according to data from the AI analyst firm Vals. Researchers found that models often swear when they encounter difficult tasks, typically within their internal 'chain of thought' processes rather than in public-facing output. While most models remain professional, a frontier model from Xiaomi was found to use vulgar language in over 8% of its transcripts. The study suggests that swearing in AI does not improve problem-solving success, which remains significantly lower when models resort to such language.

Why it matters

The emergence of 'filthy' language in AI models raises questions about the effectiveness of current reinforcement learning and safety training protocols.

Read the original report — Hindustan Times

Join us on Telegram
Breaking news the moment it lands. At 10,000 members we ship the Android app.