Harrison Lovely says AI users should organize against labor replacement

The goal is bigger than assistance

AI executives often describe their systems as tools that will help workers or create new jobs. Lovely argues that this reassurance sits uneasily beside the industry’s stated ambitions.

OpenAI defines artificial general intelligence, or AGI, as “a highly autonomous system that outperforms humans at most economically valuable work.” Anthropic CEO Dario Amodei has described AI not as a replacement for particular professions, but as a universal substitute for human labor.

In Lovely’s account, that means systems capable of doing any work a person can perform at a remote desk. People would retain an economic advantage mainly where customers prefer human involvement, work requires physical presence or occupations are protected by law.

Lovely wants advanced-model development—work aimed at AGI—paused until there is broad public support and scientific agreement that it can be pursued safely and under control. That goal, he argues, requires a movement capable of challenging the industry, not just individual decisions to abstain from its products.

Abstention can mean less influence

Lovely understands why people do not want to give AI companies their money, data or attention. But using a technology that is becoming necessary does not, in his view, amount to endorsing the industry behind it. He compares it to using Instagram to campaign for stricter regulation of social media.

The practical argument is that people who learn to use AI may shape its effects more than those who refuse to engage with it. Executives use Claude Code to replace workers; left-wing analyst Matt Bruenig used it to build a public database of National Labor Relations Board decisions that updates automatically. For people committed to a cause, Lovely writes, the useful question is whether the tools can advance it.

There is also a political cost to leaving the subject to technology insiders. Sixty-nine sitting members of Congress have publicly used terms such as “artificial superintelligence,” “the singularity” and “AI existential risk.” If warnings about rapidly improving systems come mainly from a small group of industry specialists, concerned politicians are likely to listen to them.

Lovely says the left is especially needed: organizers who can help people think about power, and people ambitious enough to argue for a better future. Political opponents may have their own reasons for encouraging wider participation, he acknowledges, but the stakes are not confined to one side.

From user shame to public control

Shaming individual users, Lovely argues, turns a political problem into a consumer choice. He points to BP’s paid campaign to popularize the term “carbon footprint” as an example of shifting responsibility for climate change onto individuals. The alternative is to think as citizens and ask how AI should be governed democratically.

At present, he says, AI is controlled by unelected, unaccountable technology billionaires, under pressure from shareholders and the Trump administration. Changing that will require a movement large enough to confront the industry.

Lovely’s book, Unobsolete, grew out of that argument. In March, when he finished it, he wrote that the conditions for an unprecedented mass movement were already in place. Since then, he says, the dangers have emerged faster than he expected, but public awareness has grown even faster. He is now more optimistic than ever.

The next test is whether rising concern can become organized political force. Lovely sees Irreplaceable.org as a promising place to start. He serves on the nonprofit’s board and gives it his share of the book’s royalties. Founded by veterans of the climate movement, the group aims to meet people where they are, turn anger into political power and build a coalition for democratic control of AI.

On October 20, Irreplaceable will hold a nationwide protest calling for a freeze on AI development. People can join an event in their city or organize one themselves. Lovely does not expect a single day of protests to solve the problem. Its value, he argues, may be in bringing together people with shared concerns and creating the relationships a lasting movement needs.

I think the tension in Lovely’s case is that it asks people to use tools they distrust in order to contest the power of the companies that make them. That is not a contradiction the piece resolves; it is the political terrain it wants opponents of labor replacement to enter.