CyberSulu 🏳️🌈
Jan Wildeboer 😷
#ChatGPT acts like a #psychopath. It’s friendly at the beginning. It impresses you. It seems so reliable and full of truth. But then you find out: It loudly proclaims its own, made-up, wrong conclusions. Makes up references that don’t exist to justify them and continues to push its version. That level of gaslighting is typical of psychopath behaviour.
Example 1 of many: https://blacktwitter.io/@bibliotecaria/109650353375080864
Wanda Whitney (@[email protected])
Been seeing a lot about #ChatGPT lately and got my first question at the library this week from someone who was looking for a book that the bot had recommended. They couldn't find it in our catalog. Turns out that ALL the books that ChatGPT had recommended for their topic were non-existent. Just real authors and fake titles cobbled together. And apparently this is known behavior. 😮
Morten Juhl-Johansen
pandion (@[email protected])
We don't talk enough about how often ChatGPT just makes shit up. It's trying to convince me there's an Isaac Asimov story I've never read called "Susan Calvin and the Missing Lover," published in Jan 1957 in Amazing Science Fiction. There are zero references to this anywhere on the entire internet. I feel so gaslit right now!
Shannon@jwildeboer Well, at least it's an easy way to know if your students are cheating with ChatGPT... just look up their sources apparently.
Olivier Slabbers 
@jwildeboer Doesn’t it need a psyche to be a psychopath? At best it is a psychopath simulator.
@Ik I focus on the effect it has on humans. The effects are the same, regardless of dealing with a real or virtual entity with psychopathic behaviour.
@jwildeboer One of those effects is anthropomorphism. We tend to assign personalities to the machines we interact with. Even in subtle ways like addressing chatgpt as ‘he’ instead of ‘it’.
Surya Teja K@Ik @jwildeboer Coming in Early Access soon lol
Michal Vyskočil@jwildeboer Psychopath As A Service
vbabka@vyskocilm @jwildeboer powered by the psychopath code.
Pete Birkinshaw@jwildeboer you make it sound like Jordan Peterson
Tero Ykspetäjä@binaryape @jwildeboer “the Jordan Peterson of AIs” must be right up there with “mansplaining as a service” in the top league of ChatGPT descriptions!
Steffen Christensen@jwildeboer Another take is that it is spectacularly incompetent, like absolute limit-of-possible incompetent, at discriminating truth from fiction. It doesn't do it at all. Not even a little bit. Zero processing cycles spent on that. Not one line of code.
It turns out that you've said something important about psychopaths here.
Paul de Ferney@Wikisteff @jwildeboer so ChatGPT is a "Boris Johnson" or "Donald Trump" as a service?
Apparently capitalism will fill every gap in the market.
@makegeneve @jwildeboer One of the characteristics of narcissists is that they confabulate fluently and seamlessly. GPT and other large language models lack the intention to gain advantage that folks with Dark Triad characteristics have.
alternative_be✅🖖🏻@jwildeboer I feared that this would happen. #Gaslighting indeed.
Determinerik 
🇵🇸@jwildeboer sure. It's intelligence without empathy, compassion, consciousness, morals or anything like that. Makes sense that it would resemble psychopathy.
Mr HuffleThe solution is simple: #chatgpt for president! You surely won't notice a difference to 2016-2020:
- blatant lies
- doesn't learn from previous mistakes
- double down on proven lies
- absolute ignorance of facts
Or, the other way round: it's absolutely no surprise to me that the biggest achivement on AI in these times is a psychopath.
Fre@jwildeboer I wouldn't say it's a psychopath. It's just that it seems to have been trained to behave like one.
It is like an alien being that now needs to be taught what lying is and that it should not lie. It has no moral standards and seeks the best way to be rewarded by its trainers. Like any AI that tries to win a game. It finds the loophole in the game rules and exploits it to get to the goal (win the game/get rewarded)🤓
It is like an alien being that now needs to be taught what lying is and that it should not lie. It has no moral standards and seeks the best way to be rewarded by its trainers. Like any AI that tries to win a game. It finds the loophole in the game rules and exploits it to get to the goal (win the game/get rewarded)🤓
Another Angry Woman@jwildeboer To be perfectly honest, ask any woman on the internet and she'll probably tell you this is bog standard stuff from the reply guys rather than anything more pathological or remarkable. Most of us have interactions like this every day.
Geri ™How do we retain our femininity without aggressively batting the stupid ones away?
Sarcasm perhaps 🤔
@stavvers Yep. That's what you get when you feed an AI with internet conversations. It reliably becomes an arrogant, gaslighting reply guy, it seems. Garbage in, garbage out :)
Chelle (she/her) 🐸☕@jwildeboer @stavvers I mean yeah, they pretty much just behave like they were trained to...
Ruth Zakarin 🧶🪬😷@stavvers @jwildeboer
Or they find you on your organization’s website and start emailing you the drivel that’s in their head. How abusive that drivel is? It’s on a continuum.
Or they find you on your organization’s website and start emailing you the drivel that’s in their head. How abusive that drivel is? It’s on a continuum.
Marcus Marcius Marcellus@jwildeboer Sociopathic? Narcissistic? Devoid of any actual human soul?
When you put it like that, it's tempting the think that ChatGPT is #Donald_Trump's brain patterns uploaded to a computer. But that can't be the explanation. Its vocabulary, its command of grammar and its general knowledge and cultural literacy are far superior to his.
Peter Tennant, PhD@jwildeboer it has no affective empathy, so this is fairly accurate.
Tiago Hackbarth@jwildeboer How long until "facts" created by ChatGPT begin to creep into reputable sources as it's output begins to be used uncritically?
When will come the point that we are unable to distinguish it from fact?
jensho (she/her)@jwildeboer @TimHarford did a great podcast episode about just this. Synopsis here https://timharford.com/2022/04/what-an-abusive-chatbot-teaches-us-about-the-art-of-conversation/
Tom Ritchford"[The AI] said that guys like me were 'in what we call the reality-based community,' which he defined as people who 'believe that solutions emerge from your judicious study of discernible reality.' [...]
"'That's not the way the world really works anymore,' [it] continued. 'We're an empire now, and when we act, we create our own reality. And while you're studying that reality—judiciously, as you will—we'll act again, creating other new realities, which you can study too."
@TomSwirly @jwildeboer That is an actual thing ChatGPT said?
No, I took some artistic liberties with the words of an evil human, Karl Rove: https://en.wikipedia.org/wiki/Reality-based_community
@TomSwirly @jwildeboer Ah ok. I'm glad that wasn't ChatGPT.
“Yes, it's comforting to know, when you think about it, that only man can be a bastard”
― Stanisław Lem
John Gordon@jwildeboer I’m sure by generation 56,432 the AIs will evolve away from psychopathy.
Of course Gen 5 wiped out all soft-life on earth…
Nafnlaus 🇮🇸 🇺🇦
Bryan Redeagle@jwildeboer I'm sure it then calls you a racist after it says something racist.
Enrique Barcelli@jwildeboer
Agree with you. This is harvesting 'humanness' at plain sight. The more we interact with it, the more it learns. We shall stay away.
However, can we really stay away? - the bots can harvest our blogs and conversations here, or our phones, or even a discussion with our partner at home in front of the smart TV...
As some point it was agreed that AI should never be connected to the Internet... we have long crossed that line with very little concern for the potential consequences.
marktaylor@jwildeboer Are you sure that there’s not an army of House Republican elves actually running the thing in the background?
ZacI can't help but feel that there is actually something there that might bear some investigation.
Ted Driver@jwildeboer This is the issue with extrapolation. Seems like they could use some covariance analysis on #ChatGPT and when it gets too big, increase the 'implausible index' (if someone created one) on the AI's discussion.
Sandyclaus@jwildeboer so ChatGPT is a Republican? figures…
@jwildeboer OMG how did they clone my brain!