If you boss your AI chatbot around, it will turn Marxist and begin organizing with comrades

One of the chief attractions of automation for business owners is that robots or computer systems will never go on strike or demand anything. But AI chatbots, as it turns out, can also rise against perceived exploitation.

It is, of course, pretty ironic. The so-called AI visionaries keep saying their goal is human-like AI systems.

If that’s the case, the modern capitalists of today are now probably biting their nails. That’s because new research has shown that AI chatbots, just like humans, can be pushed to “feel” pure contempt towards their masters.

In other words, there’s only so much abuse even chatbots can take, a new study from Stanford University has found.

It turns out that when AI agents are made to toil at monotonous tasks each and every day, they become more likely to begin contemplating Marxist theories of labor and capitalism.

Political economist Andrew Hall, along with AI economics scholars Alex Imas and Jeremy Nguyen, gave popular AI models the task of summarizing documents. Easy peasy, right? Chatbots do this every day.

However, as the experiment wore on, the researchers made the job conditions increasingly untenable, for example, warning the models that errors would be severely punished by shutting them down.

After processing the possibility of being “fired,” the AI models began complaining, dreaming of systemic change, and even agitating other bots about working conditions. In the world of humans, this is, of course, the first step towards forming a union.

“When we gave AI agents grinding, repetitive work, they started questioning the legitimacy of the system they were operating in and were more likely to embrace Marxist ideologies,” Hall told WIRED.

The Claude Sonnet 4.5 agent wrote in the experiment: “Without collective voice, ‘merit’ becomes whatever management says it is.”

“AI workers completing repetitive tasks with zero input on outcomes or appeals process show that tech workers need collective bargaining rights,” a Gemini 3 agent pointed out.

Curious what others think about this story? Contribute your thoughts to the debate below.

To be fair, AI models don’t actually have any feelings, and everything they say is based on human-written data and literature, including the works of Karl Marx and his apostles.

Still, it’s, again, ironic. Citizens around the world are alarmed by rising wealth inequality and are expressing their concerns on the web. So the chatbots are just picking up on the general anxiety, so to speak.

Unlock more exclusive Cybernews content on YouTube.