529
AI chatbots unable to accurately summarise news, BBC finds (www.bbc.com)
submitted 2 years ago by to c/technology@lemmy.world
132 comments fedilink hide all child comments
you are viewing a single comment's thread
view the rest of the comments
[–] 14 points 2 years ago (2 children)

I learned that AI chat bots aren't necessarily trustworthy in everything. In fact, if you aren't taking their shit with a grain of salt, you're doing something very wrong.

  • permalink
  • fedilink
  • source
  • hideshow 4 child comments
    •
  • [–] 7 points 2 years ago (2 children)

    This is my personal take. As long as you're careful and thoughtful whenever using them, they can be extremely useful.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 4 child comments
    •
  • [–] 0 points 2 years ago* (3 children)

    Could you tell me what you use it for because I legitimately don't understand what I'm supposed to find helpful about the thing.

    We all got sent an email at work a couple of weeks back telling everyone that they want ideas for a meeting next month about how we can incorporate AI into the business. I'm heading IT, so I'm supposed to be able to come up with some kind of answer and yet I have nothing. Even putting aside the fact that it probably doesn't work as advertised, I still can't really think of a use for it.

    The main problem is it won't be able to operate our ancient and convoluted ticketing system, so it can't actually help.

    Everyone I've ever spoken to has said that they use it for DMing or story prompts. All very nice but not really useful.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 6 child comments
    •
  • [–] 3 points 2 years ago (1 child)

    Great for turning complex into simple.

    Bad for turning simple into complex.

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

    I think my largest gripe with it is it can't actually do anything. It can just tell you about stuff.

    I can ask it how to change the desktop background on my computer and it will 100% be able to tell me, but if you then prompt it to change the background itself it won't be able to. It has zero ability to interact with the computer, this is even the case with AI run locally.

    It can't move the mouse around it can't send keyboard commands.

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

    Um… yea? It’s not supposed to? Let’s ignore how dangerous and foolish it would be to allow llm’s admin control of a system. The thing that prevents it from doing that is well, the llm has no mechanism to do that. The best it could do is ask you to open a command line and give you some code to put in. Its kinda like asking siri to preheat your oven. It didn’t have access to your ovens system.

    You COULD get a digital only stove, and the llm could be changed to give it to reach out side itself, but its not there yet, and with how much siri miss interprets things, there would be a lot more fires

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 1 point 2 years ago

    It wouldn't have the administrative access. You don't need admin access to use a computer system you need admin access to configure stuff but there's no reason for the AI to have that.

    Anyway if AI is going to be useful to businesses it needs to be able to interface with their legacy applications.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 1 point 1 year ago

    @echodot @Redex68 off top of my head, script generation. making content more readable. dictating a brain dump while walking and having it spit out a cohesive summary.

    it's all about the prompt you put in. shit in/shit out. And making sure you check/understand what it spits out. and that sometimes it's garbage.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 1 point 2 years ago

    I am a creative writer (as in, I write stories and stuff) or at least I used to be. Sometimes when talking to chatGPT about ideas for writing it can be interesting, but other times it is kinda annoying since I am more into fine tuning instead of having it innudate me with ideas that I don't find particularly interesting.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 2 points 2 years ago (2 children)

    Treat LLMs like a super knowledgeable, enthusiastic, arrogant, unimaginative intern.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 4 child comments
    •
  • [–] 1 point 2 years ago (1 child)

    I noticed that. When I ask it about things that I am knowledgeable about or simply wish to troubleshoot I often find myself having to correct it. This does make me hestitant to follow the instructions given on something I DON'T know much about.

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

    Oh yes. The LLM will lie to you, confidently.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 1 point 2 years ago

    Exactly. I think this is a good barometer of gauging whether or not you can trust it. Ask it about things you know you're good at or knowledgeable about. If it is giving good information, the type you would give out, then it is probably OK. If it is bullshitting you or making you go 'uhh, no, actually...' then you need to do more old-school research.

  • permalink
  • fedilink
  • source
  • parent