As reports of ChatGPT sending its users down dangerous mental health spirals mount, OpenAI seemingly can only think of one thing to say.
  • Electricd@lemmybefree.netEnglish
    1·
    30 days ago

    What I meant is:

    You can’t expect LLMs not to do that because that’s not technically possible at the moment

    Companies should display warning and add some safeguards to reduce the amount of time this happens