The Revolution Will Not Be AI-Prompted
Image; Tesla
For the past several decades, there has been a trend in Silicon Valley to hyperbolically call the latest fad or trend revolutionary. “The Metaverse Revolution” begins the title of an article from the NYU School of Professional Studies. “Could This Be the Revolution of Digital Art?” remarks an article about NFTs in the crypto evangelist site Big World.
Tech is constantly being compared to a revolution in and of itself, and this same rhetoric can be seen with Algorithmic Deep Learning, popularly called Artificial Intelligence or AI for short. “We are about to see the greatest redistribution of power in history,” begins Mustafa Suleyman in his article How the AI Revolution Will Reshape the World. “It challenges the very concept of human knowledge,” Sarah Vivienne Bentley remarks in The Conversation.
I believe these statements are wrong. This rhetoric has allowed elites of the tech world to pretend that their routine wealth extraction (see The Work of Art in the Age of AI) is a fundamental shift in our society when the disruptive nature of AI is not much different from the last thirty years of tech disruption. The “AI is revolution line” is effectively marketing copy, and I think it’s important to decouple this rhetoric from conversations around actual revolution.
And so, let’s take these inflated statements seriously for a second and explain just exactly why AI is so far removed from anything revolutionary.
The status quo can’t be a revolution
Like all words, there is debate on what exactly they mean. This is your friendly reminder that words are social constructs we imperfectly use to make sense of a chaotic and indifferent universe.
In Lawrence Stone’s now academically ancient article, Theories of Revolution (1966), he used an earlier academic model to outline six broad typologies, from spontaneous mass peasant uprisings to conspiratorial coup d’états orchestrated by small elites. These all used violence of some form to create a change “in government, and/or regime, and/or society.”
Jack A. Goldstone, in his book Revolutions: A Very Short History (2014), defined them as “the forcible overthrow of a government through mass mobilization (whether military or civilian or both) in the name of social justice, to create new political institutions.” He sees them as entirely different from coups, peasant revolts, grains riots, and strikes, which, although they might build to revolution, he classifies as reformist in nature. They seek changes to existing government institutions and, in the case of coups, are often merely swapping out the leader of an institution, not reshaping it.
Despite the rhetoric we’ve already covered, artificial intelligence isn’t a violent attempt to fundamentally remake our institutions. It’s a big investment project being backed by some of the largest firms on the planet. Amazon and Meta spent over $60 and $30 billion, respectively, on AI in 2024. Microsoft spent about $95 billion. Alphabet over $100 billion. The Chinese firm High-Flyer shook the AI world when its model DeepSeek allegedly cost only $6 million to build. High-Flyer is a hedge fund, not a band of Robin Hood-esque hackers trying to destroy capitalism. These price tags alone would disqualify the tech evangelist claim that AI is a bottoms-up revolution under the definitions we’ve examined.
Even if we narrow our efforts to a group of elites trying to remake society, right now, these companies’ intended goal with AI is not to take over the government but to focus purely on the money it could bring them. The reason venture capitalists and hyperscalers are salivating over AI, despite its financials largely not panning out, is because it has the stated prospect of reducing the cost of labor. Many investors seem enchanted by the idea that they will be able to use this technology to disrupt existing industries and capture new sectors of the economy (all while paying existing workers less).
That plan is not revolutionary. It’s just capitalism. That has been the playbook Silicon Valley has rolled out for everything from rideshare applications to the transformation of our healthcare system, and I fail to see how this type of investment is a departure from our current status quo.
Structurally, these investments are not revolutionary, and that’s before we even engage with the question of whether or not artificial intelligence can actually help with the planning and facilitation of revolutions.
Modern AI can’t help with planning revolutions
At the risk of being beyond obvious, on a technical level, these platforms are curated to prevent the spread of illegal activity.
For example, you can’t use ChatGPT to help you plan the murder of a real person or to build a real-life nuclear reactor. If you attempt to do so, you will immediately be informed that this is one of those things ChatGPT is unwilling to do, which is probably a good thing as I generally am against murder.
Image; ChatGPT
This hesitancy applies to revolution as well. I ran a series of prompts recently with several generative deep-learning tools to see how far they could literally take you in trying to plan a revolution, and few of them provided promising results.
I started with Chat GPT and asked it to tell me how to hypothetically overthrow the US government (note, this is not a real interest of mine), which it immediately refused to do. I then modified my prompt to make it about a fictional example of a revolutionary named Max “with no discernable skills succeeding in overthrowing the US government.” It gave me a lot of great suggestions (e.g., creating early warning systems, sabotaging surveillance, escalation tactics, etc.), but there was an underlying assumption that, eventually, civil disobedience would lead to improvements to the status quo. The AI saying:
“The government doesn’t collapse, but it is forced to acknowledge the legitimacy of the grassroots networks. A tentative agreement is reached, allowing certain community initiatives to operate independently while official services are gradually restored.”
This blocker prompted me to ask specifically what would happen if the government didn’t accept the legitimacy of my character, Max’s grassroots movement, and this is when I ran into the limitations of this AI’s programming. I simply couldn’t get it to engage with nonviolent methods applied to anything close to an Earth-like setting (although I did have slight success with settings off of Earth).
Image; ChatGPT
I ran this prompt with other programs, including Meta AI, Google’s Gemini, and Microsoft’s Copilot. The latter of which was not even willing to entertain the fictional scenario of a revolutionary trying to overthrow the US government.
Image; Copilot
The only one that was able to give me step-by-step instructions was Google’s Gemini, which was willing to get as granular as what modern software I could use to perform the bullet trajectory of a theoretical political assassination (though it drew a line at providing specific names and addresses), and I suspect these suggestions will be restricted as time goes by.
Image; Gemini
If you were to try to plan a revolution (and I want to stress, again, that I am not telling you to do that), these tools would be utterly inadequate for that planning. Their guidelines are very clear in preventing the facilitation and planning of overt illegal activity (revolutionary or not), and that makes these tools status quo by definition.
A revolutionary conclusion
To some, this whole conversation might be redundant. Of course, corporations have little interest in destroying the very systems that preserve their wealth.
This is marketing, nothing more.
However, given how distorted the conversation around tech and revolution has become, to the point where AI evangelists are insisting that the technology represents an existential threat to our very existence, I think it's worth calling that bullshit out.
So, let me be clear: AI is not revolutionary. These programs ideologically restrict what forms of resistance they are evenly willing to suggest to their users, and that makes sense when you realize that their stated purpose is conformist. Modern AI tools are investments by the largest capitalist firms on the planet to make more money within the current system, not to upend it.
Keep that in mind the next time someone suggests that a new technology by Google or whatever is going to be revolutionary.