home.social

#elizaeffect — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #elizaeffect, aggregated by home.social.

fetched live
  1. Did anyone else notice the latest argument from "AI" pushers is along the lines of "But you're questioning experts in the field!"? Classic Appeal to Authority there.

    #ai #aislop #elizaeffect #llm

  2. #AI #chatbot #LLMs #GenAI #anthropomorphization then sadly followed by #aipsychosis #aidelusion

    Originally 1966 #ELIZAeffect coined by #Weizenbaum who escaped nazi germany as teenager, to become #mit #compsci pioneer
    en.wikipedia.org/wiki/ELIZA_ef

    programmed a simple psychiatrist #chatbot —JavaScript web.archive.org/web/2025011412
    Typical answers:
    - tell me more about that
    - please go on

    80s tv recreation of his unexpected surprise once #Eliza archaic bot deceived his secretary:
    youtu.be/RMK9AphfLco

  3. @edwiebe I have some horror stories to tell about techs who made that awful and absolutely incorrect assumption.

    The people have their own lives, and many of whom you probably depend on THEIR expertise and TRUST them to do their job, be that driving TTC or grocery clerk or banker or plumber, you EXPECT they are telling the truth and you BELIEVE you have recourse to the courts if they lie.

    None of that is true about the IT industry. After 55 years in it, I often wish I had spent my time in an honest job, like pimping, or used-car sales.

    #ELIZAEffect was an accident back then, but today it is TOTALLY deliberate.

    So, perhaps you meant "People flogging IT technology are, on average, either incredibly stupid, or fantastically sociopathic."

    You could run a poll 🤣

  4. #ElizaEffect
    Have you heard about the Eliza Effect?
    This was named after the first AI computer system which was built in the 1960s.

    The Eliza effect happens when humans project human traits and abilities on machines and objects, like empathy, emotions etc.
    This explains why some humans these days falls in love with ChatGPT or have AI "friends".

    en.wikipedia.org/wiki/ELIZA_ef

    #NoAI

  5. @ai6yr

    preferred term - "powerful AI"

    Indeed. Power requires Intel. Leverage. Extortion … #elizaeffect #posiwid

  6. @CSMonitor

    Users don't want it, it can't be trusted, it gobbles the planet like a demon, and its (poor) performance per terawatt is beginning to plateau, so then why are the managerial non-users so keen to deploy it? What's the real business case?

    #ELIZAeffect + #SurveillanceState = #AGI

    #POSIWID

  7. That @cbcnews is completely ignorant of the #ELIZAeffect and where a major world-class news-team is oblivious, that does not bode well for the general population.

    AI fanboi long-promised 'singularity' is no different from any other #Rapture cult.

  8. @rupdecat from talking with people outside the Tech bubble, I learned that they don’t have enough information to understand the limitations of #LLMs. They rely on what is written in the press. People don’t know about the #ElizaEffect and that LLMs are just giving the illusion of a dialog, Even in the Tech bubble, few know about the so-called decoding procedure and the usage of sampling for constructing an LLM response. Even less are knowing how sampling works. All this needs a lot of education!

  9. #Chatbots – egal wie einfach oder komplex gemacht sie sind – nutzen den ELIZA-Effekt: eine Wahrnehmungsverzerrung, die Menschen dazu bringt, tendenziell zu glauben, was gut formuliert ist.
    en.wikipedia.org/wiki/ELIZA_ef

    Die fatale Tendenz, Chatbots zu anthropomorphisieren, wird verstärkt, wenn in diesem Zusammenhang von „künstlicher Intelligenz“ gesprochen wird. Wenn sie LLMs verwenden, dann sagt das und nicht „KI“.

    Vielleicht ist es weniger fatal, Chatbots mit Tieren zu vergleichen. @emilymbender hat dafür „stochastische Papageien“ vorgeschlagen.

    Wenn ich lese, was @lcamtuf dazu schreibt, bin ich entgegen seiner Empfehlung doch versucht, Chatbots zu vermenschlichen – als Trickbetrüger.
    lcamtuf.substack.com/p/llms-ar

    #Chatsbot #llm #bard #GoogleBard #ChatGPT #OpenAI #ELIZA #ElizaEffekt #ElizaEffect

  10. @KeithDevlin I like this statement “Humans are “hard-wired” to ascribe understanding (and a whole lot more) to any agent that interacts with us using language” - yes there are so many people suffering from the #elizaeffect, in particular if they prefer #wishfulthinking to #analyticalthinking.

  11. @rstockm Du sagst, dass du weißt, dass Midjourney ein Programm ist, aber gleichzeitig sprichst du diesem Programm eine Seele, Talent und Kreativität zu und genderst es. #elizaeffect

    Midjourney versteht nicht, was es tut. Es kann Strings zu geklauten Bildelementen und Stilen zuordnen. Aber es besitzt kein Wissen über die Welt.