Climate Activists Target OpenAI, Google, Amazon in NYC Protests

Climate protesters picketed the Manhattan offices of OpenAI, Google and Amazon this week, declaring that tech companies — not artificial intelligence itself — are the real "danger to humankind" and the planet.

The demonstrations unfolded during climate week in New York City, an annual gathering that draws environmental activists, policymakers and industry figures to the city. On Monday evening, protesters assembled outside OpenAI's unmarked Manhattan offices, the home of the company behind ChatGPT. Some held signs reading "Eat the rich, save the planet."

The protests did not stop there. Over the following two days, activist groups also picketed outside the offices of Google and Amazon, two of the largest investors in AI infrastructure in the world. The coordinated targeting of all three companies signals a shift in how climate activists frame their criticism of the AI industry.

A broader pattern of disruption

The office pickets were only one part of a wider week of action. Protesters, including clergy, were arrested while disrupting a closed-door AI health summit held in the city. The disruption of a private AI event marks an escalation: activists moved beyond public streets and into spaces where executives, researchers and investors gather to conduct business.

Tonight, the protests will shift to Brooklyn, where demonstrators plan to target a gas power plant. The plant had been set to close — until it was purchased to power AI datacenters.

That detail cuts to the heart of the protesters' argument. The activists contend that the immediate, measurable environmental cost of AI — the electricity consumed by datacenters, the extension of fossil fuel infrastructure, the water and land demands of computing facilities — poses a present danger, in contrast to the abstract, long-term catastrophic scenarios that dominate much public discussion of AI risk.

Why the framing matters

The protests come at a moment of intensifying public anxiety about artificial intelligence. The demonstrations followed the resignation of an Anthropic employee who quit his job with a warning that sharpened an already-growing unease about the technology's trajectory.

"The people building AI earnestly believe that it could kill us all by the end of the decade," the former Anthropic employee wrote, according to The Guardian's report on his departure.

That statement, from an insider at one of the most prominent AI safety-focused labs, added fuel to a debate that has moved from research circles into mainstream political discourse. The climate protesters outside OpenAI's doors this week offered a different answer to the same underlying question: if AI poses a danger, where does that danger actually lie?

Their answer is corporate behavior. The signs, the pickets and the targeting of specific companies — OpenAI, Google, Amazon — articulate a position that the responsible parties are identifiable, addressable and located in office buildings in Manhattan, not in some hypothetical future system.

The choice of targets reflects the scale of the companies' involvement in AI. OpenAI makes ChatGPT, the chatbot that triggered the current wave of commercial AI investment. Google and Amazon are among the world's largest operators of cloud infrastructure, the computing backbone on which AI systems are trained and run. All three companies are driving the datacenter expansion that has reshaped electricity demand projections.

The power plant protest

Of the week's actions, tonight's protest at the Brooklyn gas plant may carry the most direct symbolic weight. The plant was scheduled to shut down. Its purchase to power AI datacenters reversed that trajectory, converting a retiring fossil fuel asset into infrastructure serving the AI industry's growth.

For activists, the plant embodies their core claim: that the AI boom is not an abstract future risk but a present-day force shaping energy policy, extending the life of fossil fuel facilities and redirecting resources toward computational expansion. The protest there tonight will bring that argument to the physical site where those decisions materialize.

Context: a movement finding its target

The week of protests illustrates how climate activism is adapting to the AI era. Rather than addressing AI as a monolithic technology, the protesters named companies, named executives' venues and named a specific power plant. The message on the signs and in the targeting choices was consistent: the "danger to humankind" is the set of corporate decisions being made now, by identifiable actors, with measurable environmental consequences.

This positioning places the protesters partially at odds with the framing advanced by some AI researchers and industry figures, who emphasize longer-term existential risks from advanced AI systems. The Anthropic employee's resignation warning — that AI builders believe the technology "could kill us all by the end of the decade" — represents that catastrophist framing at its most alarming. The climate activists outside the offices this week did not dispute that anxiety is warranted. They redirected it.

Whether this framing gains traction will depend in part on how the AI industry's energy footprint evolves. As datacenter construction accelerates and more fossil fuel facilities are repurposed to serve it, the protesters' concrete, localized grievances offer activist networks a target-rich environment: specific offices, specific summits, specific power plants. The Brooklyn demonstration tonight suggests this strategy is just beginning.