Dr. Damien P. Williams: Magus, Werewolf, Cyborg, Bi: I literally warned y'all about exactly this kind of harmful outcome from GPT integrations, & even told you *why/how* it'd happen. And if it's showing in résumé sorting, then bet it's about break as having been happening in patient portals & police reports & lending application sorters & on & on & on

I literally warned y'all about exactly this kind of harmful outcome from GPT integrations, & even told you *why/how* it'd happen. And if it's showing in résumé sorting, then bet it's about break as having been happening in patient portals & police reports & lending application sorters & on & on & on

“UW researchers found that ChatGPT consistently ranked resumes with disability-related honors and credentials — such as the ‘Tom Wilson Disability Leadership Award’ — lower than the same resumes without those honors and credentials.”

ChatGPT is biased against resumes with credentials that imply a disability — but it can improvewww.washington.edu UW researchers found that ChatGPT consistently ranked resumes with disability-related honors and credentials — such as the “Tom Wilson Disability Leadership Award” — lower than the same...

Back in 2022, at the American Bar Association’s Labor & Employment Law Conference, I was on a panel where I discussed exactly this range of possibilities, as well as the ways ableism was gonna show up in the "AI" facial recognition surveillance interview systems that are now increasingly everywhere…

The session was called “Using Artificial Intelligence in Hiring and Recruiting: The Future is Here.” And yeah, we told 'em; it sure fuckin is.

Back in 2022, at the American Bar Association’s Labor & Employment Law Conference, I was on a panel where I discussed exactly this range of possibilities, as well as the ways ableism was gonna show up in the "AI" facial recognition surveillance interview systems that are now increasingly everywhere…

Here's some links. Y'know, for whatever. acsjournals.onlinelibrary.wiley.com/doi/full/10.... www.zinio.com/article/time... www.e-elgar.com/shop/usd/han...

Artificial intelligence and medicine: Mounting risks amid the promiseacsjournals.onlinelibrary.wiley.com <em>Cancer Cytopathology</em> is an international cytopathology journal publishing research related to topics concerning the etiology of cancer, its diagnosis and prevention.

Hmm, that sounds like a foreseeable outcome. If a company knew, or should have known, that the tool would likely discriminate... is using it anyway evidence of intent?

that headline is so, so depressing. "this thing we don't need oppresses several groups. but we can mitigate that probably." so ominous and horrible. "my flippant choice will hurt you. I'm confident I can make it hurt you a little less, though, eventually." just unbelievable. critical thought 😢

(related, I wrote this on my phone, which insists that 'oppresses' is not a word and wants to substitute in 'oppressed.' no phone, I have no need to make the subject invisible with that verb, I'm fine)

huh, my phone (ios) just suggested “oppresses” when i typed “oppress”. wonder why yours doesn’t! not that autocorrect is that consistent for me, i suppose

machine learning is a mirror that reflects our societal values back at us. people would rather think that since it’s computers that it’s perfectly fair

I saw this coming a mile away. This shit was obvious from well over a decade ago when Facial Recognition software, a type of AI, had racist biases, and when used by police was known as a tool and weapon to help maintain racial profiling. Systematic persecution maintained through computer systems.

Feel like this should lead to discrimination lawsuits.

I'm sure it will!

I've raised this stuff in convos & even made myself unpopular in some situations where I refused to back down from this. I realized the people that don't want to believe this stuff even in the face of *evidence* only PRETEND to care about equity, equality & diversity. They don't actually give a shit

Once again for people in the back: Any AI, LLM, ML, or statistical system that uses historical (or current real-world) data is going to be biased, because those biases are baked in to our society and require active correction.

Yup. www.americanscientist.org/article/bias... bsky.app/profile/wolv...

Here's some links. Y'know, for whatever. acsjournals.onlinelibrary.wiley.com/doi/full/10.... www.zinio.com/article/time... www.e-elgar.com/shop/usd/han...

Artificial intelligence and medicine: Mounting risks amid the promiseacsjournals.onlinelibrary.wiley.com <em>Cancer Cytopathology</em> is an international cytopathology journal publishing research related to topics concerning the etiology of cancer, its diagnosis and prevention.

Bias Optimizerswww.americanscientist.org AI tools such as ChatGPT appear to magnify some of humanity’s worst qualities, and fixing those tendencies will be no easy task.

There’s been a major example seemingly every month for years. Who could have predicted it’d still be true?! www.bu.edu/articles/202...

Algorithms Were Supposed to Reduce Bias in Criminal Justice—Do They?www.bu.edu Data can discriminate, says BU’s Ngozi Okidegbe, the first dual-appointed professor to the School of Law and the Faculty of Computing & Data Sciences

Mad they won't let me load gifs here, because we clearly need this one. (I was trying to post it to another topic earlier today, so had it handy.) It's early to call it, but AI is gunning to be one of the biggest trashfires of the 21st century.

We’re not even a quarter of the way through this century and the candidates for trash fire of the century are really stellar.

I’ve spent half the morning shopping for just a toaster. Appliances ratings are just a trash fire. How can I have confidence in AI when they can’t even build a decent countertop toaster?

Another day in the life of Capitalism F*cks Everything Up.

Who could have predicted that a materialism engine would view the world in exclusively material terms?

If the programmers are biased, the robots will be too.

Even if the programmers aren’t biased, if the training data is the robots will be.

This has been around before the “AI Boom” and it’s absolute BS. they claim these videos aren’t biased but you don’t know what arbitrary criteria it’s based on. the fact in some instances they’ve made allowances for ND people shows it’s BS for everyone

let's take a deep delve into everything chatGPT has ruined

It's so much easier to list the handful of things it seems to actually have good uses for.

how long do you have

@charleswlogan.bsky.social

Do not use!

At present, any "AI" will embody all the prejudices and biases of existing data. While a human can learn to be conscious of their own biases and make a conscious effort to overcome them, what we call "AI" now cannot. We can only create in our own image.

I work in the employment sector, and have seen a number of startups promising to "automate hiring", every time I wait to see how they re-invented racism/ableism/sexism/etc this time.

Why I NEVER complete that info on the application disclosures for jobs.

Post