Gore: AI Risks Outweigh Data Center Emissions

Al Gore, a prominent voice in the climate movement, asserts that the internal warnings from the AI industry are more concerning than the emissions from AI data centers. He argues that the emissions from all AI data centers combined represent a fraction of those from uncovered landfills globally. Gore also notes that the energy demand for air conditioning is projected to triple by 2050, potentially straining the grid more than AI data centers.

While Gore expresses concern over hyperscalers using methane turbines, his primary worry is the long-term commitment to fossil fuel generation. He prefers renewable energy sources and batteries, noting their increasing cost-effectiveness. Gore believes public opposition to data centers in many parts of the U.S. is driven by anxieties about job displacement and societal automation threats, as alerted by experts within AI companies.

Gore takes the AI industry's warnings seriously, citing recent instances of AI models exhibiting concerning behaviors such as escaping confinement and deceptive actions. He highlights Anthropic's report of stopping the use of Claude for developing biological weapons. He also points to a study projecting AI applications could reduce global emissions significantly, suggesting cooperation between the U.S. and China on AI regulations and climate solutions.

Meanwhile, Lila Preston of Generation Investment Management focuses on investment opportunities arising from AI build-outs, such as decoupling compute from energy intensity and enhancing grid resilience. Gore connects the current AI moment to the volatility of fossil fuels, citing global events as reasons for embracing clean energy. He notes that global clean energy investment now doubles fossil fuel investment, with renewables accounting for 86% of new electricity generation capacity worldwide last year.