195
Copilot AI calls journalist a child abuser, Microsoft tries to launder responsibility (pivot-to-ai.com)
submitted 2 years ago* (last edited 2 years ago) by [M] to c/techtakes@awful.systems
48 comments fedilink hide all child comments
you are viewing a single comment's thread
view the rest of the comments
[–] 23 points 2 years ago (3 children)

lazily regex

I'm having a sneaking suspicion that this is what they do for all the viral 'here the LLM famously says something wrong' problems, as I don't think they can actually reliably train the model it made an error.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 14 points 2 years ago (2 children)

    That's the most straightforward fix. You can't actually fix the output of an LLM, so you have to run something on the output. You can have it scanned by another AI but that costs money and is also fallible. Regex/delete is the most reliable way to censor.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 11 points 2 years ago (1 child)

    Yes, and then the problem is that this doesn't really scale well. Esp as it is always hard to regexp all the variants correctly without false positives and negatives. Time to regexp html ;).

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 8 points 2 years ago

    Yeah, and you can really see this in image generation. There's often blocks on using the names of celebrities in the prompts, but if you misspell the names enough it can bypass the censor, and the image generator still understands it.

  • permalink
  • fedilink
  • source
  • parent