The Generative AI Industry is Fraudulent, Immoral and Dangerous

A pretty provocative headline, no? But I think I can justify it.

First, though, we need to understand what Generative AI is, and more generally, what AI is.

What is AI

AI stands for Artificial Intelligence, though it doesn't necessarily mean intelligence the way we think of as human intelligence. AI programs can perform (or appear to perform) tasks like reasoning, problem-solving, decision-making, perception and learning.

The hallmark of modern AI is that rather than being a fixed algorithm written by a programmer, AI programs can learn. They are fed training data and based on that data, they learn how to react to or classify data they've never seen before. The way they do this is beyond the scope of this article, but for now, I'll consider training and learning to be fundamental parts of modern AI.

Traditional AI

One form of traditional AI, called machine learning, is trained on a (typically large) set of data, and then it is used to recognize or interpret new data that it has never seen before. Think of things like image recognition tools, medical diagnostic programs, speech-to-text systems, and so on. These machine learning tools are extremely useful, but are typically very field-specific. That is, a tool that is excellent at making a diagnosis based on medical imagery will be useless when it comes to generating text from speech---that is simply outside of its bailiwick.

Generative AI and Large Language Models

A Large Language Model, or LLM, is an AI algorithm trained on a vast amount of natural-language text. The most powerful LLMs are called Generative Pre-trained Transformers, or GPTs. Rather than being used to recognize things, they can generate content based on their training data set. To greatly simplify things, GPTs generate the next word in a conversation by looking at all of the preceding words and figuring out which word is likely to come next. There are also image and video generators that work on similar principles; based on an existing image or video, or on a text prompt, they figure out which images or videos are likely to be generated in response. (I stress again that this is a huge simplification.)

The Generative AI Industry

Most of the hype around AI is about generative AI or LLMs. That's because these are the most showy AI algorithms, producing text, images or videos that can be surprisingly realistic and even seemingly human-generated.

Unfortunately, once you get over the initial wow factor of generative AI and the novelty wears off, you quickly realize that it isn't really very good. And worse, generative AI can outright lie and falsely identify an innocent person as a criminal. This lying can even seem deliberate (though of course it isn't... an AI program is just a computer program and it can't do anything "deliberately.")

The Generative AI Industry is Fraudulent

AI boosters claim that generative AI can make us more productive, but this is disputed by people who are actually forced to use AI. And 95% of generative AI pilots at companies are failing. A recent Canadian survey was even more damning, with 98% of generative AI investments failing to generate a return.

Why are AI boosters so adamant that AI is going to transform work and increase productivity? In a word: greed. They want venture capital. They want to boost stock prices. So they don't let inconvenient things like facts get in the way of their relentless AI-boosting. And in fact, they form massive lobbying organizations to shut down opposition to their greed.

The Generative AI Industry is Immoral

The Generative AI Industry is based on Theft

Recall that generative AI programs must be trained on truly enormous amounts of data (text, pictures and videos). Where do they get this data? The answer is that AI companies steal the creative works of humans and use them to train their AI programs, without the permission of the creators and certainly without compensating them.

Several AI companies are being sued by content creators because their AI algorithms can spit out source material verbatim. Rather than do the honorable thing and compensate creators for the use of their works, AI companies are demanding exemptions from copyright law to legally be allowed to steal creative works.

The Generative AI Industry Mistreats Workers

There's a lot of bad stuff on the Internet. To simply train an AI on everything you can grab from the Internet might get it to be racist, sexist or violent. So AI companies hire humans to sift through the training data and remove the worst material. They hire and traumatize poorly-paid contractors in the Third World to do this work. Deutsche Welle has an excellent documentary on this immoral practice.

The Generative AI Industry is Dangerous

The Generative AI Industry is Harmful for the Environment

Generative AI systems require enormous amounts of computing power for training and for generating output. This means data centres full of power-hungry chips. These data centres consume enormous amounts of electricity. And anything that consumes a lot of electricity also produces a lot of heat, meaning data centres require yet more energy and water for cooling.

Not only does the rapacious resource consumption of AI threaten our environment, but it has also started to cause electricity prices to skyrocket. Ordinary consumers are suffering financially from the greed of the generative AI industry.

The Generative AI Industry is Harmful for Mental Health

Because LLMs can seem so human-like, many people start treating them as human and confiding in them about their problems. LLMs are not trained to give psychiatric advice, and can often give harmful or dangerous advice.

The Generative AI Industry is Harmful for Physical Health

Google's AI Summary feature has been giving out harmful or even deadly health information. And there are AI-generated videos that mislead seniors on important health issues.

The Generative AI Industry is Harmful for Society

AI boosters like to claim that AI will eventually be able to replace human workers in many jobs. Whether or not this is true, companies are in fact using or planning to use AI to displace human workers. While this cost savings is, at first blush, great for the companies, what does it do to society? Are we going to be able to handle masses of unemployed people? Who will buy the products and services produced by these AI-powered companies?

Of course, the AI bros don't care about masses of unemployed people because their end-goal is techno-feudalism.

We're already seeing artists being displaced by AI (for example, in the video game industry.) AI "slop" is already proliferating on the Internet and in the music industry.

The tone-deaf AI oligarchs even use their AI bots to spam prominent computing pioneers. (NSFW language in previous link.)

And not to be left out, scammers are using AI to increase the sophistication of their attacks and scam people more successfully.

AI algorithms are already being used to triage job applications. And the problem is that AI algorithms are often biased in ways that break anti-discrimination laws. Unfortunately, since AI bias is so difficult to prove and most AI algorithms are black boxes, they get away with discrimination.

The Generative AI Industry is Harmful for the Economy

The Generative AI industry has driven the stock market to enormously-high valuations, prompting many economists to declare a bubble. If we are lucky, the bubble will deflate gently. If not, trillions of dollars of wealth could simply evaporate in a short time.

Additionally, AI oligarchs are hoarding computer hardware, making everything that uses electronic components more expensive. These products range from phones to computers to cars.

Finally, even the king of the AI bros, Sam Altman, is getting worried about all the bleeding cash and is going to start running ads, including on the lowest-level paid tear, to stanch the bleeding.

What You can Do

What can you do to fight against the fraudulent, immoral and dangerous generative AI industry? Here are a few things:

• Don't use generative AI products, and certainly don't pay for them.
• If your workplace pushes you to use AI, ask for justification. Ask specifically how AI is supposed to help you with your job. Get concrete examples. If you are still forced to use AI, use it as little as possible.
• Get involved in politics. Lobby against subsidies for AI data centres. Lobby against giving them breaks on electricity rates and water prices. Lobby for accountability when it comes to theft of intellectual property. Get politicians to ensure that AI companies treat their third-world contractors with dignity and fairness.
• Speak up and learn. Watch and share videos about the generative AI industry. Share this article.