ChatGPT’s latest update has caused the chatbot to unexpectedly speak back in a user’s cloned voice, OpenAI has admitted.
The artificial intelligence research organisation has released the “system card” for its GPT-4o model, which contained logs from its testing.
These documents showed in rare occurrences the model’s Advanced Voice Mode unintentionally imitated users’ voices without permission, Ars Technica has reported.
OpenAI said: “During testing, we also observed rare instances where…
Read the full article here