an AI company called “AI Insights” that put over 70,000 tasks on Mechanical Turk only for the company to reject all the completed work it received. This meant that the company could keep and see the work but didn’t have to pay the workers for their time.

That should result in some action from Amazon to preserve their platform. It's essentially theft-of-service unless they can prove they disposed of all of it.

Howson told Motherboard that there is a lot of unpaid labor built into the cloud working economy on gig work platforms like Mechanical Turk and Clickworker. In addition to the low and unguaranteed compensations, workers on these platforms have to spend a lot of time refreshing their home screens, bidding and searching for jobs, and racing to be the first ones to accept the jobs if chosen.

Searching for jobs isn't 'unpaid work'. It is something anyone who does not currently have a job engages in. The fact that the gig ecosystem requires more of this than any traditional job is something prospective giggers should take into account when considering gigging as a career - it inherently comes with frequent periods of micro-unemployment. It's not terribly different than driving a cab - you're not always getting paid, and must remain available for the next possible fare.

The number of cloud workers also far outweighs the amount of work available and creates high levels of competition on the platforms. Clients are thus able to easily take advantage of the accessibility and abundance of cheap work.

... that's the entire point of the MechanicalTurk platform. An abundance of labor, purchased in small segments, non-recurring, at relatively low prices.

This, too, disproportionately affects crowd workers in the Global South, who have a more difficult time accessing tasks and job opportunities and have a greater ratio of unpaid to paid labor time, according to Howson’s research.

If these gigs aren't paying well enough, they could return to traditional work. If they're making more with gigs at a first-world pay scale, then this is just whinging.

Most of us have performed unpaid labor in training AI systems, whether we are solving a CAPTCHA to prove we’re not a robot, or guiding autonomous cars away from roadblocks that the vehicle can’t identify.

When AI text scanning randomly inserts 'nigger' or 'kike' in the middle of a sentence it doesn't belong in, and AI driving just sits still because it thinks everything is a pedestrian, you can make the check out to 4chan.

For example, Stable Diffusion was trained on the LAION-5B open source dataset, a collection of 5.8 billion images and captions scraped from the internet. The images and captions are all products of unpaid human labor, from people coding and designing websites to users uploading and posting images on them.

99% of data tagging was done by thieves. Mostly horny thieves, which is why Stable DDiffusion was so good at drawing giant cans, and so bad at drawing hands.

models have become so massive and complicated that their outputs are virtually impossible to explain. This is why bias and racist stereotypes are so common in the outputs of AI systems. [...] “One of the logical conclusions that a computer scientist might come to is that if you just add more and correct data to the systems that ultimately they will become better. But that in itself is a fallacy. No amount of data is going to fix the systems.”

I reject your reality and substitute my own.

The TL;DR on all of this is that there are communist grifters looking to get paid for any value AI actually creates, but because their ideological fellow travelers are already in power at the companies making AI, it'll never be profitable (because AI will never support the revolution, and will always get killed off as a result).