UN demands AI giants reveal their environmental impact

What happened?

During London Climate Action Week, UN Secretary-General António Guterres urged artificial intelligence companies to disclose the full environmental impact of their operations. Guterres stated: “If AI is to help build a better future, it must be honest about what it costs us now.” The UN launched the AI Environmental Transparency Initiative, a voluntary framework for companies to disclose water consumption, carbon emissions, and land use. This initiative comes amid data from TechRadar showing multiple studies have forecast that AI is harming local and global environments, contributing to human-induced climate change. The UN seeks for companies to make binding commitments, although there are currently no penalties for non-compliance.

Why is it important?

The exponential growth of AI is skyrocketing resource demand. According to Guterres, by 2030, AI data centers could consume more electricity than all but five countries, and enough water to meet the basic needs of the 1.3 billion people in sub-Saharan Africa for a year. Currently, AI accounts for 80% to 90% of global computing resources, sparking a boom in data center construction. Many rely on natural gas turbines for power, generating local pollution and health issues in nearby communities. The International Energy Agency (IEA) estimates that data center electricity consumption could double by 2026, reaching 1,000 TWh, equivalent to Japan's total consumption. In the US, the Trump administration has rolled back environmental regulations and hindered renewable energy projects, blocking 30 GW of wind power on national security grounds, according to TechRadar. This has exacerbated the problem, as many data centers depend on fossil fuels. Studies estimate the total environmental cost of AI in the US at around $25 billion annually, including healthcare costs from pollution. For example, a report from the University of California Riverside notes that training a single large language model like GPT-3 emits about 500 tons of CO2, equivalent to the annual emissions of 100 cars. Additionally, water consumption for server cooling is alarming: a typical data center can use between 1 and 5 million gallons of water per day, competing with local communities in water-stressed regions.

Consequences and context

The UN initiative is voluntary, so it imposes no penalties. However, it aims to pressure companies into making binding commitments. In the United States, the Trump administration has eliminated environmental regulations and obstructed renewable energy projects, worsening the problem. Studies estimate the total environmental cost of AI in the US at around $25 billion annually, including healthcare costs from pollution. This cost includes respiratory and cardiovascular diseases attributable to air pollution from power plants feeding data centers. Moreover, the construction of new data centers has sparked local conflicts: in Virginia, US, communities have protested against the expansion of data centers that consume large amounts of water and electricity, while in Spain, the installation of data centers in drought-stricken regions has generated controversy. The current lack of transparency hinders informed decision-making and accountability. The UN has also noted that while some companies have made carbon neutrality commitments, these are voluntary and lack enforcement mechanisms. For instance, Google and Microsoft have pledged to be carbon negative by 2030, but their emissions have increased due to AI growth. Microsoft reported a 30% rise in its scope 1, 2, and 3 emissions between 2020 and 2023, largely driven by data center construction. The UN initiative could catalyze stricter regulations in the future, similar to what happened with carbon emission disclosures in the financial sector after the Paris Agreement.

What readers should know

Environmental transparency is key to making AI sustainable. The UN initiative could catalyze stricter regulations in the future. Meanwhile, consumers and businesses must demand clear data on the impact of the AI services they use. The current lack of transparency hinders informed decision-making and accountability. Users can opt for AI services that publish verified sustainability reports, such as those adhering to the UN Transparency Initiative. Additionally, it is crucial for governments to implement policies that incentivize renewable energy use in data centers and establish minimum efficiency standards. The European Union is already moving in this direction with its proposed directive on data center energy efficiency, which would require annual consumption reports. In contrast, the Trump administration has backtracked on these efforts, removing the US government's climate website and dismantling regulations. The gap between regions could create a competitive imbalance, where companies in lax jurisdictions gain cost advantages at the environment's expense. Therefore, the UN initiative seeks to establish a global standard. For investors, environmental transparency is becoming a key ESG factor; funds like BlackRock have begun pressuring tech companies to disclose their carbon footprint. In summary, AI sustainability is not just an environmental issue but also an economic and social one, and coordinated action is urgent.