Moonshot, a leading AI developer, announced new safeguards in its latest Kimi model to curb its ability to generate hacking tools. The move follows growing concerns that advanced language models could be weaponised for cyber‑attacks, a risk highlighted by recent high‑profile breaches. By restricting code‑generation functions that facilitate exploit development, Moonshot aims to balance innovation with responsible use.

Industry analysts see the restriction as a signal that AI firms are taking proactive steps to mitigate misuse, especially as governments tighten regulations on AI‑driven cybercrime. While the limitation may temper some developer enthusiasm, it could set a precedent for broader industry standards on model safety and responsible deployment.

This story was written by a language model from reporting by The Economic Times. Check the original before quoting it.