Email or username:

Password:

Forgot your password?
Marcin Wichary

Has anyone written about how textual generative AI feels strangely close to toxic masculinity in some respects? The absolute confidence in everything stated, the lack of understanding of the consequences of getting that confidence wrong for important questions, the semi-gaslighty feeling when it “corrects” itself when you call it out on something. It so often feels like talking to someone one would despise and avoid in “real life.” I’m curious if anyone did some writing on this.

26 comments
Jef Poskanzer :batman:

@mwichary I've heard it called mansplaining as a service.

Michael B. Johnson

@mwichary not sure about “writing”, but that’s definitely the commentariat response I have seen to it online for my memory of it existing.

i.e. “so… it’s like the internet”?

etc.

Julius Welby

@mwichary After pondering this piece I suddenly blurted “Oh my god! They’ve modelled Brian!” (Name changed). We’ve probably all had to work with or under a Brian at some point.

relsqui

@jwelby @mwichary this is a much more compelling version of my gut response which was "of course it is, look who's designing it"

pettter

@mwichary I've heard it described as "mansplaining as a service", but I'm not sure about any long-form writing.

Madagascar_Sky

@mwichary
That sounds more of 'Privileged Upbringing' set of the 'Unearned Confidence' attitude. Unless the LLM also then asked for your number and told you 'that a smile wouldn't hurt'.

mORA

@mwichary Yesterday I was talkin’about AI and used the hashtag #AIsplaining

(In English is «With AI we will all have answers to questions we haven't asked.»):

mastodon.uno/@mora/11248468537

NanoRaptor

@mwichary Gods yes! I was telling a friend only in the last month it sounds like a sales guy I used to know who would overconfidently state a wrong thing and I'd just have to bring in that one fact that showed it up "oh! of course of course yes, sorry about that, anyway <another wrong thing>"

jwz

@mwichary HAL confidently incorrecting an increasingly-agitated Scarlett Johansson.

Jonathan Irons

@mwichary This is a bit of a tangent, but I recommend this piece on how gen AI works in a similar way to psychic mediums. softwarecrisis.dev/letters/llm

Colin

@mwichary is this one of the reasons why it’s so exhausting to read?

I swear I can pick up on this within a sentence or two (if other elements on the site didn’t give the game away already).

Mathieu Perona

@mwichary The statistician in me suggests that these statistical engines do just what they are designed to do: mirror the main features of their training data set.

Colin Rosenthal ❤️🇮🇱❤️

@mwichary That's why it's been dubbed "mansplaining as a service" (first known usage @cwebber I believe). Other epithets are available.

DELETED

@mwichary Good summary of what seems to be the general feelings around it all.

Petra van Cronenburg

@mwichary Yes, there are studies about the social and ethical impact of biased #AI #LLM #generativeAI, especially in questions of masculinism, racism or homophobia. It's a fact that the popular models are trained mainly by men (with the #SiliconValley "philosophy"***) on men dominated content. The latest is this #UNESCO study: cepis.org/unesco-study-exposes
This test became quite well-known in 2023: rio.websummit.com/blog/society

#bias #biased #GenderBias #gender #AIEthics

@mwichary Yes, there are studies about the social and ethical impact of biased #AI #LLM #generativeAI, especially in questions of masculinism, racism or homophobia. It's a fact that the popular models are trained mainly by men (with the #SiliconValley "philosophy"***) on men dominated content. The latest is this #UNESCO study: cepis.org/unesco-study-exposes
This test became quite well-known in 2023: rio.websummit.com/blog/society

Petra van Cronenburg

@mwichary *** And it starts with the profession itself: how many women work in AI tech and how are they treated? The newest information: salon.com/2024/05/21/coercive-

You can search for more studies.

Nini

@mwichary Well yeah, if there's one thing that embodies LLMs it's speaking complete fictions with the unearned confidence of a mediocre white man.

Faintdreams

@mwichary when the a large set of the conversational corpus of LLM's is taken from places like Twitter and a reddit, then their ... Flavour of discourse is the output

Go Up