Arthur T Knackerbracket has processed the following story:
AI is great for rich and powerful people and for tech giants trying to boost profits. Otherwise, artificial intelligence and the automation it enables can be harmful, nonprofit Mozilla concluded in a report published Monday.
"In real life, over and over, the harms of AI disproportionately affect people who are not advantaged by global systems of power," Mozilla researchers conclude in the 2022 Internet Health Report. "Amid the global rush to automate, we see grave dangers of discrimination and surveillance. We see an absence of transparency and accountability, and an overreliance on automation for decisions of huge consequence."
[...] But Mozilla doesn't like the fact that Big Tech funds a lot of academic research and that relatively few papers -- especially among those most widely cited -- focus on AI's social problems or risks.
Among Mozilla's suggestions are new laws. "Regulation can help set guardrails for innovation that diminish harm and enforce data privacy, user rights, and more," Mozilla said. Also on Monday, Mozilla released a five-part podcast on its concerns about AI.
Reference: 2022 Internet Health Report
(Score: 5, Insightful) by Thexalon on Thursday July 21 2022, @11:12AM (3 children)
The good news is that what is alleged to be AI isn't as correct as a lot of people think it is. It's relatively easy to confuse it beyond all recognition if you put some effort into it. For example, I've gotten advertiser targeting to think I, a cisgender dude, am a good target audience for tampon ads.
The bad news is that what is alleged to be AI isn't as correct as a lot of people think it is. Which means that those with access to the outputs will treat it as accurate even if it's not and falsely believe that it's objective when it isn't.
I always thought of AI algorithms in roughly the same terms as the disclaimer that's been on polls on this site and on the Green Site for decades now: "This whole thing is wildly inaccurate. Rounding errors, ballot stuffers, dynamic IPs, firewalls. If you're using these numbers to do anything important, you're insane."
The only thing that stops a bad guy with a compiler is a good guy with a compiler.
(Score: 2) by c0lo on Thursday July 21 2022, @01:44PM (1 child)
Wait, don't tell... mmm... your hobby is supersonic muzzle velocity tampons [youtube.com], amiright?
https://www.youtube.com/watch?v=aoFiw2jMy-0
(Score: 5, Funny) by RamiK on Thursday July 21 2022, @01:53PM
If you use "cisgender" in your google search, you're in the market for tampons one way or another.
compiling...
(Score: 4, Insightful) by JoeMerchant on Thursday July 21 2022, @02:00PM
> those with access to the outputs will treat it as accurate even if it's not and falsely believe that it's objective when it isn't.
I'll just leave this [theguardian.com] here.
Україна досі не є частиною Росії Слава Україні🌻 https://news.stanford.edu/2023/02/17/will-russia-ukraine-war-end
(Score: 0) by Anonymous Coward on Thursday July 21 2022, @01:44PM (4 children)
Google, OpenAI, and the rest are gatekeeping AI under the guise of preventing harm. We need open source algorithms that aren't being hoarded by Silicon Valley giants.
(Score: 3, Interesting) by c0lo on Thursday July 21 2022, @01:50PM (3 children)
Very likely, those algos already are open source.
Good luck to you in:
* clustering together zillions of compute units to run those algos. No your old smart phones aren't enough.
* collecting and curating a training data set (that's not pure garbage) to train those algos
https://www.youtube.com/watch?v=aoFiw2jMy-0
(Score: 2) by The Vocal Minority on Friday July 22 2022, @05:16AM (2 children)
Pre-trained models are available which can still be used to do cool things, including fine-tuning for better performance on task specific datasets. So in some ways the heavy lifting only needs to be done once.
My understanding is that GPT3, much like the BERT models, would be very amenable to pre-training if it was publicly available.
(Score: 2) by c0lo on Friday July 22 2022, @06:45AM (1 child)
And how does this makes the providers (of the pre-trained models) act less of an "AI gatekeeper"?
Context: the OP was ranting about
https://www.youtube.com/watch?v=aoFiw2jMy-0
(Score: 2) by The Vocal Minority on Friday July 22 2022, @05:36PM
By "available" I mean on places like github or huggingface under permissive licences. Yes the situation is a little more complex than the original rant would have us believe, but the argument is not entirely without merit, and your objection doesn't nullify it entirely as I have pointed out. Unfortunately I don't have time to address the complexities at the moment, and I don't count myself an expert on this topic anyway.
(Score: 5, Insightful) by JoeMerchant on Thursday July 21 2022, @01:57PM (2 children)
Information tools generally, or at least often, play in zero-sum-games. Unless the AI is eliminating waste or inefficiency, it will be used to get more profit for its owners at the expense of those providing the profit. This is no different than all the wetware business analysts that have been working the trades for the last 500+ years.
Stock market analysts, tax attorneys, product marketing specialists - they're all focused on getting more profit for their employers, and it comes at the expense of those not employed by the same people. Replacing, or more often: augmenting, these parasites with AI only enhances their abilities, it doesn't change them in any fundamental way.
Україна досі не є частиною Росії Слава Україні🌻 https://news.stanford.edu/2023/02/17/will-russia-ukraine-war-end
(Score: 3, Interesting) by RamiK on Thursday July 21 2022, @04:42PM (1 child)
The difference is that neural nets operating costs scale exponentially so the statistical odds of small players surviving in a data-driven market where AI tools are available becomes abysmally small.
compiling...
(Score: 2) by JoeMerchant on Thursday July 21 2022, @06:59PM
The big players will be careful to leave enough crumbs for the rest to avoid significant political backing for antitrust violations, or not and pay the price. This translates roughly to them keeping the bulk of the gains of new technology for themselves and others' getting a fraction of their gains, which is little different than every other technological revolution in the last 200 years, if you think about it.
Україна досі не є частиною Росії Слава Україні🌻 https://news.stanford.edu/2023/02/17/will-russia-ukraine-war-end
(Score: -1, Troll) by Anonymous Coward on Thursday July 21 2022, @04:26PM
Perhaps if "the vulnerable" would pull themselves up by the bootstraps and stop sponging off the rest of us, then they might amount to something.
As Better.com CEO Vishal Garg wrote when he fired 900 employees, “HELLO—WAKE UP BETTER TEAM. You are TOO DAMN SLOW. You are a bunch of DUMB DOLPHINS and…DUMB DOLPHINS get caught in nets and eaten by sharks. SO STOP IT. STOP IT. STOP IT RIGHT NOW. YOU ARE EMBARRASSING ME.”