jisp

games

Home Projects Blog Archive
GenAI Misconceptions | jisp

GenAI Misconceptions

January 17, 2026

Hello! I am jisp, and I work in tech.

I feel like there are a lot of online misconceptions around generative AI (“genAI”), so I wanted to set a few things clear:
  1. No, generative AI is not the same technology as machine learning or AI used in other applications.
    • Generative AI is fundamentally different than a lot of other uses for machine learning and other artificial intelligence use cases. The issues lie with the mass training that the software uses. The amount of data trained for generative AI is so indescribably large, it tells us that the datasets that generative AI are trained on:
      1. contains people’s works without their permission.
      2. require excessive energy output to process
    • The next few points will go in detail why these are problems. These issues are unique to generative AI that is being sold to the public by companies such as OpenAI, Sora, and Midjourney.
  2. Yes, generative AI will cause devastating harm to our environment and our society
    • This technology requires data centers to train, generate, and share data to users. Major tech companies have gone around the desires of the ordinary person to push generative AI onto us.
    • Cornell University estimates that in 2027, AI data centers around the world will have a water withdrawal of over 4.2 billion cubic meters. They are intentionally trying to hide this!
    • Microsoft’s data centers can lead to the evaporation of 700,000 liters of clean freshwater today. The idea that a “closed loop” system works perfectly is silly. That’s now how matter, heat, and energy works. If any of our energy or water systems worked without externality, we would not have this conversation.
    • Emissions from Google have increased by 50% between 2019 and 2024, mostly from data centers. The International Energy Agency estimates that data centers will double their energy use between 2022 and 2026. This rapid acceleration of burden on the world’s energy grid results in disruption of access, loss of life, and an increase in consumer costs.
    • Data centers being built now are anticipated to use the same amount of energy as large cities! In most scenarios, the financial cost is forcibly passed on to consumers.
  3. No, generative AI is not “inevitable”
    • In technology historian Dr. Mar Hicks’ interview with FastCompany, they discuss how framing technology as inevitable is a powerful tool for those in power to shift the conversation from whether or not something works well to instead the mitigation of harm for this “inevitable” thing. This is an intentional tactic that weakens any and all criticism! Hicks’ next statement for me, seals the deal. They say: “This is a technologically deterministic way of thinking, because it produces this idea that technologies shape society when, of course, it’s usually the other way around. It’s society that creates and chooses what those technologies are and should be.” I think this is a great way to look at it!
    • The concept that generative AI is inevitable is a hoax, illusion, and a lie that venture capitalists are trying to sell the entire world. We don’t have to buy into it.
  4. No, generative AI is not the savior of the working class
    • The idea that generative AI is the technology that will uplift the working class and bring us Universal Basic Income (UBI) is misinformation. This belief comes from the idea that we can separate technology from the systems in which it is created. That isn’t true. Generative AI is created by and for the capitalist class. It’s conception and basis is developed around the idea you can steal from existing works faster than society can catch up to you or laws can punish you for it. Generative AI is predictive, it exists due to the uncredited effort of billions of people. It does not create new works; it generates a mashup of what came before and diffuses the credit.
    • For example, Midjourney is a text-to-image software. Midjourney is used for generating images, including use as logos for companies and businesses. This software was trained on a dataset of over 5billion images on the internet. This dataset, provided by Laion, is scraped off of the internet. Laion claims they are a non-profit research organization, but if Midjourney is trained on the datasets that Laion have scraped from the entire internet, then the hundreds of millions of dollars that Midjourney recieves from users and VC investors is theft. Midjourney is just one of the many growing examples of these mass-theft machines that are being built and packaged as artistic tools for all! In reality, it is an illusion of creation wrapped over shameful exploitation of all.

Acerola has a great 33 minute video about generative AI, I’d recommend it!

Acerola: “Generative AI” is not what you think it is

Thank you for reading !