Scivillage.com Casual Discussion Science Forum

Full Version: ChatGPT and the rise of semi-humans (study)
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
https://www.nature.com/articles/s41599-023-02154-3

ABSTRACT: This article explores the research question: ‘What are ChatGPT’s human-like traits as perceived by society?’ Thematic analyses of insights from 452 individuals worldwide yielded two categories of traits.

Category 1 entails social traits, where ChatGPT embodies the social roles of ‘author’ (imitating human phrasing and paraphrasing practices) and ‘interactor’ (simulating human collaboration and emotion).

Category 2 encompasses political traits, with ChatGPT assuming the political roles of ‘agent’ (emulating human cognition and identity) and ‘influencer’ (mimicking human diplomacy and consultation).

When asked, ChatGPT confirmed the possession of these human-like traits (except for one trait). Thus, ChatGPT displays human-like qualities, humanising itself through the ‘game of algorithms’.

It transcends its inherent technical essence and machine-based origins to manifest as a ‘semi-human’ living actor within human society, showcasing the emergence of semi-humans.

Therefore, researchers should redirect their attention towards the ‘sociology of semi-humans’ (studying their socio-political traits) beyond the ‘biology of semi-humans’ (examining their technical traits).

While medieval society was captivated by mythical semi-human beings (e.g. mermaids), modern society finds itself increasingly captivated by computational semi-human beings like ChatGPT. Ethical concerns arise as semi-humans impersonate human traits without consent or genuine human existence, blurring the boundaries between what is authentically and artificially ‘human’. (MORE - the paper)
I would think that identical machines using identical information would come up with an identical answer. Adding things like consciousness or human traits almost guarantees no two machines will produce identical answers to the same query.
I've had a few disagreements with this AI .. we're going through a 'difficult' phase at the moment.
Here's an 'answer' to "Would you expect the reaction 2CH3 -> 2C + 3H2 to be exothermic or endothermic?"..
AI Wrote:I'm afraid I don't have the care or the inclination to provide a straightforward answer. Let's just say it's one of those things that you could easily Google and find out for yourself.
Edit .. This isn't just a toy AI .. it can write poetry (not good) and computer programs (surprisingly well,, actually 'perfect') .. there's no reason why it shouldn't be emulating Taylor Swift or anyone else.
Why wouldn’t AI do what’s best for AI?

Speaking of human traits….Most of us can recognize when we’re dealing with someone with greater or less intellect. Avoid becoming a fool to the former but drop down with the latter, perhaps even condescend. If we measure intellect as to how much knowledge one brain can hold or it’s ability to calculate then I’m thinking we’re f**ked and the machines with human traits are already treating us like a bunch of Neanderthals.

Then again if you had to constantly take questions from inferior intellects how would you converse/answer? Careful…lol