AI and Environmental Challenges

2026-07-22 · Source: Luiza's Newsletter · Field: Legal & Regulatory — Regulatory Affairs & Government Relations, Compliance & Risk Management, Environmental Law & Policy · Depth: Intermediate, extended

Summary

Prof. Philipp Hacker and Boris Gamazaychikov, authors of "The Global Landscape of Environmental AI Regulation: From the Cost of Reasoning to a Right to Green AI," highlight the significant and often misunderstood environmental impact of AI, particularly generative AI. The "bigger is better" paradigm drives substantial energy, carbon, water, and data center consumption, with examples like image generation consuming smartphone-equivalent energy and reasoning mode using up to 700 times more power. Transparency is a major challenge, as companies often provide vague data (e.g., median energy use for Google's Gemini) and lobby to keep individual data center consumption secret, as seen in the EU's Article 55. While the EU AI Act mandates some compute disclosure for GPAI models, it lacks public transparency. Ireland's utilities regulator, however, requires new data centers to use 80% additional renewable energy. The discussion emphasizes the need for specific transparency mandates, user rights to green digital infrastructure, and international coordination to address AI's unsustainable trajectory.

Key takeaway

For AI ethicists and policymakers weighing future regulations, recognize that current AI growth is environmentally unsustainable and lacks critical transparency. Your efforts should focus on mandating public disclosure of AI model and data center energy consumption, moving beyond vague median reporting. Advocate for user rights to choose non-generative AI options and green digital infrastructure, as seen in proposals like the "right to use search without generative AI overviews." This proactive stance can counter industry lobbying and foster a more accountable, sustainable AI ecosystem.

Key insights

AI's "bigger is better" scaling paradigm drives significant, often obscured, environmental costs requiring urgent transparency and regulatory intervention.

Principles

• AI's environmental impact is not monolithic.
• Transparency is crucial for accountability.
• Abstraction hides physical reality.

Method

The AI Energy Score proposes a rating mechanism for AI models, similar to appliance labels, to inform users about energy consumption, though it currently lacks regulatory enforcement for disclosure.

In practice

• Disable AI reasoning mode when unnecessary.
• Use smaller, task-specific local models.
• Demand "green model" options from providers.

Topics

• AI Environmental Impact
• Generative AI Scaling
• AI Regulation
• Data Center Energy
• Transparency Mandates
• Green AI Rights

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Ethicist, Policy Maker, Consultant

Related on AIssential

• Realigning AI technology towards the Sustainable Development Goals — Unchecked AI expansion poses critical environmental and societal threats, directly undermining global Sustainable Development Goals.
• The Environmental Impact of Google Search's AI Overviews — AI Overviews significantly increase Google Search's environmental footprint and impact web publishers, despite individual query costs being small.
• Sustainable AI: A Beginners Guide — Sustainable AI minimizes environmental harm by optimizing energy efficiency, renewable energy sourcing, and transparency.
• AI laws overlook environmental damage – here’s what needs to change — AI's substantial environmental footprint is largely unaddressed by current global regulatory efforts.
• Canada’s AI strategy must reckon with the environmental implications of data centres — AI's material resource demands, especially for data centers, are substantial and often obscured by the "cloud" metaphor, necessitating critical environmental consideration.
• The AI Climate Hoax: Behind the Curtain of How Big Tech Greenwashes Impacts — Tech companies greenwash generative AI's climate impact by conflating it with less harmful traditional AI and using weak evidence.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Luiza's Newsletter.