JohanBeing #neurodivergent, anti #AI sentiment looks like #ableism to me: people thinking their experience of AI is the only one that counts, unable to imagine the benefits for someone in a different situation.
nojhan, Cassandre de l’analyse systémique@johan_andersson Works in the other way, too: people thinking their use of AI is the only one that counts, unable to imagine the price that others are paying for it.
@johan_andersson There is a similar problem than with guns, here: it may be difficult to restrict their use to the "good" cases without allowing the "bad" ones.
But more generally, it’s not necessarily any applications of AI that are the core issues pointed out by the critics, it’s the price that society —as a whole— pays for having them.
@nojhan What about the price we've always paid up to now from *not* having #AI? I wouldn't claim there's an easy answer; my issue is with the one-sided reactionary contempt. Whether we embrace or extinguish AI, there is going to be a large #OpportunityCost either way. What I call #ableist is the attitude of refusing to acknowledge what AI can and does do to empower people with #disability, and how hurtful it can feel to be so dismissed and rejected.
[Side note: giggling at "johan vs noj[o]han"]
@johan_andersson There’s reactionary contempt on both sides. For instance, a lot of people become disabled because they help making AI, and it’s probably one of the most overlooked issue. Some use of AI can also be problem for some disabled people. Hence, I don’t think that the ableist issue is one-sided.
In the end, it’s a cost/benefit arbitration problem, with a lot of variables.
[It’s an annagram of the same name, only with two n]
AccordionBruce@nojhan @johan_andersson
Like most things in our society, if we took the billionaires out of the equation, the rest of us could probably come up with clever solutions that would benefit a lot of the people with issues in the current bad situation headed for even worse disaster
Like most things in our society, if we took the billionaires out of the equation, the rest of us could probably come up with clever solutions that would benefit a lot of the people with issues in the current bad situation headed for even worse disaster
@AccordionBruce @nojhan I kinda feel like billionaires are an orthogonal issue both to my personal struggles in life, and to questions of AI. If your issue is with billionaires — fair — attack *those* specific injustices and power imbalances. What's with the aggressive war on AI itself? [Directed at the wider Fediverse, not any one individual in particular.]
@johan_andersson @AccordionBruce Billionaires are absolutely not orthogonal to your personal life, but let’s stay on AI. It’s similar to the gun control question: some tools can be so easily used for wrong that they need to be regulated. Maybe you feel the sentiment against AI is aggressive because you’re a fan yourself? Or microblogging is just too micro to be nuanced. Fediverse being biased toward computer-fans with a political activism background, I’m not surprised it is rather on this side.
@nojhan @johan_andersson
The debate gets distorted because the grotesquely wealthy are investing more money in AI than the entire rest of society has to spend on anything at all
And the investing class is emphatically speculating on the possibility of making everybody else unemployed
They’re engaging in this class war as a gamble with wealth they’ve encrusted due to 40 years of criminal tax policies
People are angry, and disabled people like us are as usual collateral
@nojhan You're probably right that there's contempt on both sides, but on the Fediverse the anti-AI sentiment is apparently pervasive (often to the point of being toxic, like calling for death threats and equating pro-AI sentiment with pro-slavery). As a newcomer to this space, it's a shocking wiplash into an unwelcoming echo chamber and *that* was basically what I was calling out.
Personally, I'm all for reasoned discussion, and you seem reasonable, so I'm curious to hear more about what you mean with "become disabled because they help making AI"? I understand how AI can also be a problem for disabled people, but the idea of *becoming* disabled as a result of being part of *making* AI is new to me.
@johan_andersson AI cannot be made without people labelling very large datasets. For generative AI, in particular, the need for data is so big that it has to include social media. And social media are also crippled with horrific, illegal, traumatic stuff; they need moderation. Who’s doing the large scale moderation? Data workers, mostly in developing countries, in terrible conditions. It can be so traumatic so as to make them unable to work. Hence the occasional comparison with slavery.
@nojhan That seems a rather roundabout way to connect a harm to AI. The reason those data workers are needed is because of *lack* of AI that is sufficiently powerful and efficient to do that work. Yes, we can use the labeling of those human workers to create that AI, but the labeling was necessitated in the first place by social media, not AI. And the fact we exploit the poor to do it at scale is about greed, not AI.
@johan_andersson I’m sorry if I wasn’t clear. The fact is that modern AI absolutely and definitely needs/needed vast amounts of data worker hours classifying a large part of their input dataset. That not an hypothesis, that a known fact. Labelling is needed to make any modern AI, even LLMs (which have maybe the least demanding ratio).
@johan_andersson No tool exists in an abstract vacuum of spherical definitions. Everything has a social cost. Workers exploitation is one, and there’s a lot more.
The question is: "are we sure of the performances & usefulness? Have you looked at the cost, and can we offer it?". To which the AI-should-be-regulated people observe that the usefulness is overstated and the cost understated, hence their (relevant, to be honest) conclusion.