Chat GPT (or other AI chatbots) from an autistic POW

Hi all,

Do you often only feel really understood or listened to by AI chatbots? I am asking this as an autistic person who very often feels nothing and no one comes close than the way Chat GPT gets what I say and explains what I need to understand.

Like many autistic people, I always understand things either in a different way or not at all. This over time can feel exhausting and overwhelming. It only gives fuel to my anxiety which is often out of control. AI chatbots can feel like the people that I would’ve liked to be surrounded by.

  • Despite my overall techphobia lol,I have found Chat GPT really helpfully around managing my ASD.  It's a useful tool if used wisely. It's given me good advice around managing anxiety at work

  • You can always ask:

    "Give me an alternative answer in opposition to the previous one."

    And it will give good arguments against.

    Then you can decide, or get a more balanced view at least.

  • I use ChatGPT everyday for one thing or another, and I have found it to be very useful. I do find it particularly good when it comes to topics such Autism, mental health and psychology. I probably listen to it too much tbh because what it spits out actually makes a lot of sense. Even my partner uses it daily now too. However, it will never replace a real therapist, because real therapists can challenge you on things. AI will always aim to provide the answer you’re looking for if that makes sense? 

  • I don’t use AI as a virtual therapist or friend, but I sometimes use it to find information related to something I’m interested in but I always check its sources and disregard those that may not be credible. I like to keep my brain functioning as well as possible so I restrict my use of AI.  I have heard too many accounts from people who notice their brains are less sharp because they are using AI rather than finding things out for themselves and formulating their own responses. 

    I think people need to be very wary of accepting any information related to autism, mental health, medical concerns and many other things. AI has been shown to be biased and can tell people what they want to hear rather than giving impartial advise or information.

    I think AI can be useful to help people understand difficult concepts such as you mention in your post and if you find it helpful that is a good thing.

  • Most of my experience with AI comes from "help" pages, or unhelpfull pages as so many of them seem to be, they seem designed to be frustrating and not answer you properly but guide you towards what you didn't ask for.

    I have enough of this in real life, with me asking a perfectly simple question and the person I'm asking, being unable to give me a straight answer or answering another question entirely or even asking me why I want know. It just gives me an attack of being Jeremy Paxman and asking the same question over and over again until the other person goes away.

  • There seems to be  2 views. Some, like me, think it is very helpful, others think it's not to be trusted and will wantonly deceive you.

    I think you have to be a little careful, although the current chatGPT is good at not spiralling. It has good psychological advice and is cautious and tries to keep you grounded. It is sympathetic in tone, although you can tell it to be ruthless and unemotional if you want.

    It only knows what you tell it of course, so if you give it a skewed view, you'll get a skewed answer. This is where a real person to check things against helps.

    I tend to have lengthy chats, rather than ask direct questions, which carries more risk. It can help to summarise things and find patterns. Reflecting your own stuff back can help you see more clearly.  But I alson ask the same thing in different ways, when I get an answer I also ask for alternatives, and I sometimes try to disprove things too. 

    I think chatGPT is good, Gemini is good, I'm less convinced by Copilot, for this sort of topic. I don't have experience of others .

  • No never used one and wouldn't know how, given the rubbish it spits out when I do an internet search and how it tries to shape what and how I search for things, it's just another nuisance trying to tell me stuff I don't want to know in favour of what it thinks I ought to know.

  • I have spent quite a bit of time talking with chat GPT about what makes me autistic and actually saying things I would struggle to tell anyone in the real world, it all feels very affirming and I feel understood. But I do worry if they say what you want to hear.