Loukas (They/Them) 🏳️⚧️In the same way as "the cloud is just someone else's computer" - "AI decisions are just someone else's prejudices".
Hey, if you like things I say, you might like my recent article about #autism, self-diagnosis and liberation 💚 https://freedomnews.org.uk/2024/02/29/autism-and-anarchy-self-diagnosis/
Ángela Stella MatutinaCrystalized into code, multiplied by billions, made into "common sense", monetized, sold, bought and buried.
Tim Ward ⭐🇪🇺🔶 #FBPE@Loukas That's what labelling is about - AI decisions (eg about whether a credit card transaction looks dodgy) are based on the "prejudice" of actual factual determinations. You might regard the victim of a fraud reporting it as a fraud to be prejudiced against fraud, but that's about as far as you can go with "prejudice".
Woodrow Douglass@TimWardCam @Loukas Almost -- AI decisions are based on the "prejudice" of determinations the labelers made, whether they're factual or not.
@bigfishrunning @Loukas "Prejudice" is about "pre-judging" something that hasn't happened yet. Labelling is about recording something that has happened.
@TimWardCam @Loukas does that mean that every labeller is an unbiased observer, and only labels things factually? Seems kind of incredible to me.
@bigfishrunning @Loukas I was trained as a mathematician, and it only takes one counter-example to disprove "every", so I never (ha ha!) claim "every".
@bigfishrunning @Loukas But fraud detection is about stats, not about "every". You're trying to do two things:
(a) be good enough at detecting actual fraud to get regulatory approval that you're trying hard enough
(b) keep false positives, and thus the cost of call centres and pissed-off customers, down.
So target driven in both directions.
@TimWardCam @Loukas fraud detection is only a very small part of the pie; the original thesis is that the output of a given AI system carries the Biases (perhaps overzealously using the word "Prejudices") of the input data and the imperfect-by-definition human labelers; I think that thesis holds true.
Nicola Romanò@bigfishrunning @TimWardCam @Loukas this reminded me of this excellent (and scary) piece on welfare fraud risk calculation in Rotterdam https://www.wired.com/story/welfare-state-algorithms/
Bornach@bigfishrunning @TimWardCam @Loukas
I love how Google DeepMind tried to compensate for those biases and ended up overdoing it in their Gemini model
https://youtu.be/Fr6Teh_ox-8
I love how Google DeepMind tried to compensate for those biases and ended up overdoing it in their Gemini model
https://youtu.be/Fr6Teh_ox-8
Magnus Ahltorp@TimWardCam @bigfishrunning @Loukas But you don’t care if one subset of your customers gets unfairly hit by your algorithms, you only care about the mean. What is the difference between this and prejudice?
@ahltorp @bigfishrunning @Loukas I'm not aware of that happening. But I work on the nuts and bolts of the engineering, I'm not a data scientist.
@TimWardCam @ahltorp @bigfishrunning it's well-documented and I encourage you to inform yourself about this aspect of your job.
@Loukas @ahltorp @bigfishrunning I attend the talks and briefings our data scientists give, thanks very much.
Kagan MacTane (he/him)@TimWardCam You're using the etymological fallacy[1] to (try to) claim that the word "prejudice" doesn't or can't mean things like "an adverse opinion or leaning formed without just grounds" or "an irrational attitude of hostility directed against an individual, a group, [or] a race". But it can[2], and it's *quite clear* that @Loukas and @bigfishrunning meant it that way.
1: https://en.wikipedia.org/wiki/Etymological_fallacy
2: https://www.merriam-webster.com/dictionary/prejudice, sense 2
Steve Leach@Loukas Being as we are in an election year and a bizarrely polarized and dis-informed political climate, this immediately made me think:
"Polls are just an average of everyone else's prejudices".
@Loukas Electoral/ML comparisons are often on my mind though. The Electoral College? That's Max Pooling.
Ariaflame@stevenaleach @Loukas Oh not everyone's. Judging by the way they do polls it's an increasingly unrepresentative sample.
Patrick Morris Miller@ariaflame @stevenaleach @Loukas Polls only measure the opinions of people who actually answer the phone when it doesn't recognize the number.
@kentenmakto @ariaflame @Loukas I was referring to the polls as in elections, polls as in surveys are more problematic - how do you reach the average person in the modern world? Most of us don't have landlines, and I know I personally don't answer calls from unknown numbers. But yea, analogies to bias in sample selection apply to polling more directly.
@stevenaleach @kentenmakto @Loukas It's also a problem when you have a country where voting is not considered a duty or required and so many people are apathetic.
@stevenaleach @Loukas
A computer model that predicts the outcome of an election based on polling data is not all that different from what generative AI is doing -- predicting what a human would write in response to a prompt, or predicting what an image would look like to match a given caption.
The outcome can both be accurate as well as undesirable.
Steve Hersey@Loukas
Well put!
It's worth expanding that to the readily evident "the cloud is just someone else's computer, and they don't care what happens to you or your data so long as they profit from it," with the obvious analogy to AI decisions.
Dieu@Loukas the amalgamation of the internet's prejudices, which luckily has a fabulous track record of being right about everything.
Louis Ingenthron@Loukas "AI Decisions" seems like an oxymoron... AI can do a lot of cool stuff, but "make decisions" is not one of them.
Tito Swineflu@Loukas The cloud is someone else's EXPENSIVE computer and AI decisions are someone else's ENERGY HOGGING prejudice. fify
Khleedril@Loukas They're someone else's ignorance.
Eric Schutte@Loukas @martinlentink Remember, AI does NOT stand for Artificial Intelligence but for Artificial Information.
The results of AI are a statistical hallucination and NOT real information, NOT the truth and surely NOT wisdom.
And do remember how much energy and water is wasted with every AI question and answer.
Think about it, is it worth it ? Climate change will cause a war on water, be careful not to waste water.
#statisticalhallucination #AI #ClimateChange
The results of AI are a statistical hallucination and NOT real information, NOT the truth and surely NOT wisdom.
And do remember how much energy and water is wasted with every AI question and answer.
Think about it, is it worth it ? Climate change will cause a war on water, be careful not to waste water.
#statisticalhallucination #AI #ClimateChange