you are viewing a single comment's thread
view the rest of the comments
[–] [S] 25 points 2 days ago (1 child)

The funniest part is that Musk built Grok to be his supposedly fearless truth machine, and now even his own chatbot is making his social circle look like a moderation failure. At some point the problem isn’t the AI being too woke. It’s the company its owner chooses to keep.

  • source
  • hideshow 1 child comment
  • [–] 13 points 2 days ago

    AI's find patterns in data, so to be useful they require "mostly" accurate data. The only way to reliably shape them into promoting disinformation is to train them on more disinformation than not.

    In short, there is no way to make Grok stop calling Elon out for being a nazi, without poisoning it to the extent it's completely useless as an LLM.

  • source
  • parent