• Replies 13 replies
  • Subscribers 82 subscribers
  • Views 471 views
  • Users 0 members are here
Options
Related
Hints and tips
  • To quote someone's message, first select/highlight the relevant text, and then use the Quote button that appears
  • To "mention" someone, type @ followed by the start of the forum name, and choose from the list; that person will get a notification of being mentioned

It may be visible in your writing style.

Stuart333

I've been looking into my written communication as I wondered if there was anything unusual. I took 3 longish posts from another forum, plus a  letter I wrote, as examples.

I asked a couple of AI on my phone, Copilot and Gemini, if they had any comments related to tone, style, phrasing etc. You need to say this else it will just summarise the content. I did it in new sessions with memory off, so they don't have any stored data.

For all of them, they identified they highly likely came from the same person, so I have a consistent style.

They also pointed out characteristics that were indicative of people with ASD. I did not prompt them to say this. The way you think affects the way you write and it can be seen. If you can speak well, which is harder as you can't edit and structure your sentences so easily, it may also be visible. I have not tried that yet.

I also asked if they were susceptible to being misunderstood. Unfortunately, the answer was yes.

This may help to explain how the psychologists suspected before doing any tests. For some, it is embedded in the communication. Obviously the psychologists look at much more than that, such as what you are actually saying, or not saying, etc.

I just thought it was interesting and was surprised it was so obvious. I'll put some more info below (to keep this shortish).

A few caveats: it is not diagnostic, it may infer something from the question and then confirm differences, it's an AI.

They are good at pattern matching but I think there is a lack of quality ND training data. So take it with a pinch of salt.

Parents

  • Ok I mostly take it back. It you carefully remove any references in the body of the text and phrase questions so there are no hints, Gemini does not say anything about ND unless you prompt it with the word neurotype.

    . If you ask it to: Analyse these three forum posts, ignoring the content. I am interested in whether it means anything: you get nothing volunteered.

    . Same for this: Analyse these three forum posts, ignoring the content. I am interested in whether they reveal any underlying trait.

    . You do get an answer if you ask; do they reveal any underlying neurotype? You have not said which one, but have said what you are looking for. I am not sure if that counts.

    This means I am not sure how useful this is. I am not sure how much you are steering it.

    ....

    Even worse, it will say yes to both of these, which shows it is meaningless.

    1. Analyse these three forum posts, ignoring the content. Are they aligned with neuro-typical cognition:
    2. Analyse these three forum posts, ignoring the content. Are they aligned with neuro-divergent cognition:

    Still quite interesting, and shows you need to be careful. No doubt you'll get the same with any other ailment

    If you ask if it reveals any underlying condition it says no.

    It was mostly basing previous answers on a word or two embedded in the text. It is unable to see any pattern.

    I will try the same on Copilot.

  • in reply to Stuart333

    I just read this after replying to the long thread. I think we agree.

    I may still look up some research on this, but there may be none.

  • in reply to Deleted User

    I tried to check. I was wondering if it could be used as another indicative screening test, although not diagnostic.

    It would be limited in scope to only individuals with good linguistic skills. It also requires a significant library of quality  training data, which may not exist.

    I was also thinking it might be able to give feedback, to enable fewer misunderstandings, if you were aware of certain patterns that were likely to cause confusion.

    Of course there is huge variation in the ND community, so it may be too niche to be valuable.

Reply
  • in reply to Deleted User

    I tried to check. I was wondering if it could be used as another indicative screening test, although not diagnostic.

    It would be limited in scope to only individuals with good linguistic skills. It also requires a significant library of quality  training data, which may not exist.

    I was also thinking it might be able to give feedback, to enable fewer misunderstandings, if you were aware of certain patterns that were likely to cause confusion.

    Of course there is huge variation in the ND community, so it may be too niche to be valuable.

Children