Let’s have a frank conversation about only the biggest buzzword in tech right now—AI. More specifically, generative AI, because it’s all the rage right now, everything from the next piece of corporate vaporware to the next thing to ruin a profession of service workers, creatives, and all coding as we know it.

AI has dominated conversation of tech for a long time, but no matter who talks about it, it’s always divisive. You’re either shaking your fist at that AI cloud in the sky or you’re a tech bro praising the almighty robot and praying in the hope of Roko’s Basilisk.

I can’t help, but feel there’s much more nuance to the issue of AI. People will be quick to blame our culture for the extremism, but I think this sphere of tech has a bigger problem: the personification of AI.

Personifying AI removes accountability and responsibility from the entities that create it. Despite claims of valuing safety in generative AI, the larger tech firms and companies will use this to maintain their hegemony.

Today, 1 year into the AI craze, I want to cast vision on what we can do to navigate AI, how we can discuss it intelligently, and why we should advocate for AI that serves people first.

Corporate Control

Let’s start off with the one that looms over everything in generative AI—corporate influence. Every big tech company wants to get their greasy tendrils into generative AI because they believe they can use it to get more money.

Money is an easy thing to blame, but big tech controls much more in AI than we’d like to admit:

OpenAI, one of the largest AI players in the AI industry, has 49% of their for-profit stake owned by Microsoft. This not discounting for Apple’s push for OpenAI by bribing them with exposure.

While perceived to be behind, Google’s no slouch despite what media outlets and public perception is. Google has used the vast amount of data they harvest to power their AI models, including things most of Big Tech can’t achieve. This includes in video generation (Veo), real-time audio/image recognition, and strong contribution to open source AI projects.

The AI Creep

The biggest problem with corporate AI, much like OpenAI’s ChatGPT and Google’s Gemini, is privacy and surveillance. It’s no shocker that every single Big Tech company has struggled with privacy. When Google tells you that they care about your privacy, you know in your heart that privacy and Google go about as well as peanut butter and tuna fish.

The important thing to remember here is surveillance capitalism, where access to your data is sold as a commodity more valuable than the actual money. Even if you don’t willing provide information to or use these tools, your data will be fed into them because all of these companies have a stranglehold on the majority of the internet’s contents that we all use.

AI Won’t Steal Your Job, But Companies Sure Will

hidden=""

An important thing about AI is acknowledging that human rights are valuable, including privacy. There are some tasks that absolutely need to be replaced of AI because these tasks are just unsuitable to be done by human beings. Nobody wants to work in a retail warehouse. Workers are strictly monitored and kept on a strict track to pack and load things to/from a boarding area. Wouldn’t it be great if we had robots do this instead so people could just operate them instead?

Most people would balk at a statement like this, but if companies are just going to fire you and replace you of AI, maybe you shouldn’t have been working with that kind of job to begin with. You might be better at working someplace else if you’re just going to be abused in the workplace when your employer is just going to replace you with an artificial intelligence unit.

In a different universe, your job would have been replaced with something else because that company clearly thinks your job is worthless. Great example is Google’s major software development layoff (No Paywall) this year before Google I/O. It’s no different than any other kind of “-ism” in the workplace.

This is the final goalpost of artificial intelligence and the growing trend of artificial general intelligence (or the dreaded AGIs). Companies prioritize money, certain types of work, and building the “everything” machine. Sound familiar? A machine that does everything is in effect a disposable human being. Even when said machine is a glorified mimicry writer, raising money towards this reality will advantage corporations in power and disadvantage those with little.

It’s important that we acknowledge that some people have certain jobs because of their privilege or background. The problem is not LLMs; LLMs are merely tools and the cat is out of the bag at this point. It’s now about what we are going to do protect people caught in the crossfire.

Generative AI is not the end of civilization. We can’t judge it as intelligence, we judge as software. Right now, generative AI is a part of that idealistic fantasy corporations want to sell.

All the Reach, No Way to Leave (Enshittification)

hidden=""

Being subject to the whims of advertisers means these new AI platforms will face the same twisted market incentives that allowed Big Tech to gain the position of oligarchs these companies hold today.

The other problem is lock-in. Many companies safeguard the data of their models carefully and refuse to answer any questions about it.

Even the former interim CEO/now CTO of OpenAI, Mira Murati, refuses to answer to Joanna Stern of The Wall Street Journal.

STERN: What data was used to train Sora?

MURATI: We used publicly available data and licensed data.

STERN: So, videos on YouTube.

MURATI: I’m actually not sure about that.

Joanna Stern and Mira Murati, “OpenAI Made Me Crazy Videos—Then the CTO Answered (Most of) My Questions” 10:36

You can’t control many models in web-based services. Big tech knows that the open model is the most successful. A leaked internal document at Google warned that smaller, more manageable, and free AI tools would become the norm in computing.

Accountability & Accessibility

In order for real safety mechanisms to be put into place, the process would need to involve real people, politicians, and the rest of the world. The proprietary nature of most mainstream generative AI programs means everyone who isn’t part of these companies is locked out of making it safer.

To prevent this, generative AI should be interoperable and democratized. It should be made available to everyone and allow its users to whatever they want with it, the same way we would use a hammer. Because of harm that can be done, the same real world consequences and protections need to be put into AI as well.

When there are no safeguards in place, AI will become more regulated by governments, which is a good thing. What’s not a good thing is when said regulation removes agency and power from people.

As much as flack as they get, Facebook has been at the forefront of pushing people to adopt open source solutions to AI. While Google has their Gemma models, their Gemini models by contrast are closed source. Artificial intelligence needs to be widely available for everyone

Free Expression in AI

No matter where you fall on the political spectrum, we’ve seen the disastrous results of legal processes that threaten our freedoms, whether it’s copyright law or allowing big tech to self-regulate. If current trends continue, generative AI will only perpetuate what’s happening in the real world in the digital one.

People need these tools because they can often empower someone who lacks certain abilities to invent or achieve something they wouldn’t otherwise.

People who remix music gets hit the hardest by copyright. On YouTube, playing more than 8 seconds of audio is enough to get your video demonetized. YouTube doesn’t do this because they want to, but they are beholden to the laws of the land, including copyright.

Remixing music is a very complex area to address, but if done right, a remix can make a certain song or multiple songs into a song nobody expected or has ever heard before.

There needs to be protections in our own legal system that protects things like Fair Use or fair dealing because people who remix music don’t get rich. They’re normal people like you and me. And when they get hurt, we do too.

I Cheat… Because I’m Smart!

When I was younger, I considered myself a fairly good student, except in one area: math. All the way from middle school to college, I never got anything higher than a D on a math test. When I was in my college calculus, I intentionally bought the teacher’s edition of the book, just so I could copy the answers and barely pass the class.

Today, I’m not surprised when people use generative AI to write their essays or do their homework. Schools are just broken and disrespectfully ask their students to learn things they may never use in real life. If I were a student today, you bet I would be asking something like Microsoft’s Phi-3 to solve my algebra homework or calculus proofs.

Abilities Are Personal.

But hang on, is this advocating for cheating? Yes, I advocate for cheating, but I ask you: if you’re going to use AI in schoolwork, you better be smart about it.

It’s no coincidence that AI may have unintentionally made writing better. Lots of organizations from colleges to video game companies have been caught or accused of using generative AI to write things.

When I cheat by having AI do algebra, that’s not something mathematicians need to accomplish their work, it’s just labor. Mathematicians still need to come up with systems to identify patterns and craft new new theories or ideas. AI cannot do that effectively and it’s a lie that big tech continues to sell.

Writing and art on the other hand is different because it’s inherently personal. Coding and math are not personal; it’s just a bunch of numbers and letters. This is the real difference that people need to accept.

Your Work and AI Isn’t

I’m expecting some angry license-enslaved programmers to lose their minds, so let me clarify this a bit. A mindset too many people allow themselves to get stuck in is their work is important and it is.

An easy trap people get stuck in is valuing their work over their skill. If you’re going to let yourself be defined by something like your coding work or your art, you should instead pride yourself on the fact your mind is better at these things than most other people.

This doesn’t excuse you in copying work from a generative AI. Generative AI, given enough data, just repeats the corpus of information it’s been fed in an intelligible way. If you’re going to write a paper purely using generative AI, you’ll get caught real fast.

On the other hand, researching a topic can be very painful for some people. Imagine feeding something into an AI summmarizer instead. It might help someone better understand what they are reading or point the reader to key points. AI is a tool and it’s no different from how people would use SparkNotes or CliffNotes.

We don’t value art because of the intricacies of a paint brush. We don’t gush over the verbs in an author’s writing. We value art, writing, or coding because of the skill of an individual or a group of people.

Abuse Is Inevitable

Let’s speak into a fear that needs to be said: abuse of AI systems. Everything from AI-generated porn, presidential deep fakes, and theft of all “publicly available data” from every big company and budding AI startup.

The Threat to Free Expression

Let’s return to writing again, specifically to journalism. Here we see how an industry is adapting or in some cases, merely perpetuating what was always happening. Here’s some examples:

News outlets with lots of reach or big parent companies like The New York Times and Condé Nast (No Paywall) (owner of The New Yorker, Vogue, Wired) now want a piece of the AI money pie. In the case of Wired, articles like “Perplexity Is a Bullshit Machine” are written to get people on their side.

The dangerous side of cases from The New York Times and Condé Nast is the rammifications on the aforementioned freedom of expression and abuses our broken copyright system. Even if The New York Times or Condé Nast were to win their lawsuits, they prove that the copyright system is too easily abused because anybody who remotely reports information from another new outlet, is a violation of copyright.

Vilification in Vain

News reporting, even if it’s a summary of another news outlet, is protected under Fair Use. Equivalent laws like fair dealing exist in Canada, Australia, and the United Kingdom. Other countries are more gray like China and Japan.

Learn more about what qualifies as Fair Use from Stanford University’s Copyright and Fair Use Center

What’s more, if the Times and Condé Nast have their way, they aren’t really solving a problem. They have only joined OpenAI, Perplexity, or the newest AI boogeyman of the week in weakening journalism. Big outlets will be able to afford this protection, but smaller outlets and independent journalists will not.

Demonizing generative AI will only hurt everyone and innovation. What we need to aim for is democratizing generative AI. When we have interoperable, open source AI, everybody wins.

On “Public” Data

It’s worth pointing out that web scraping should be allowed and no Wired, a robots.txt file is not a sufficient means to block robots nor should they be obligated to listen to it. The Wayback Machine doesn’t listen to robots.txt on military/government webpages for preservation purposes.

People deserve the right to do what they want to publicly available information. It’s overreaching regulation like the Computer Fraud and Abuse Act that threatened web scraping in the past.

That being said, you’re not going to get my data. In my position as a creator who uses a platform that gives all my data to Google (they use my videos for training AI per the terms of service) and writes on a public website, I block all artificial intelligence bots from scraping my content. The only way the behavior of these companies ends is if you cut off the source of their data. It means discovering open source alternatives that won’t harvest your information for neural networks or line their pockets.

Closing

The bottom line is your skills are personal. Generative AI is not and nor is your work, whether it’s creative or technical. If we lose sight of this, we lose our freedoms and the little guy is hurt.

We can’t stop the abuse that Facebook, OpenAI, Apple, Google, and countless others have committed. Instead, we need to make use of these tools, specifically generative AI that’s open-source, interoperable, and respects human dignity.

If people don’t learn to adopt AI, it will fall by the wayside for normal people and become controlled by corporate interest. We’ve seen what happens the failure of open-source software in the automotive industry, virtual reality headsets, and creative software. If there isn’t enough manpower to work on the free, private, interoperable generative AI, it will fall by the wayside and remain inferior to its closed ecosystem counterparts.

Everyone needs open source AI today and I am an AI believer. Unfortunately, our laws have a long way to go and favor the powerful. I just might not use it very much because it’s still just a glorified echo machine.

Resources:

Referenced

Track Listing