Artificial Intelligence and Environmental Impact: Moving Beyond Humanizing Vocabulary and Anthropocentrism

Area of Science:

• Environmental Science
• Computer Science
• Bioengineering

Background:

• Artificial intelligence (AI) applications in digital health and bioengineering have substantial environmental footprints.
• These impacts stem from AI's high energy demands, carbon emissions, and water usage for data centers.
• The environmental costs of AI are often overlooked, despite climate change concerns.

Purpose of the Study:

• To analyze the environmental impacts of AI, particularly large language models in medicine.
• To challenge the perception of AI as purely immaterial and highlight its material ecological consequences.
• To advocate for a shift away from anthropocentric narratives in AI development.

Main Methods:

• Commentary and innovation analysis of AI's environmental footprint.
• Review of AI's resource consumption, including energy, water, and rare minerals.
• Critique of AI's associated vocabulary and its anthropocentric framing.

Main Results:

• AI's environmental impacts include significant water consumption, carbon emissions, and rare mineral extraction.
• The humanizing language used for AI masks its profound ecological consequences.
• Current AI development often prioritizes human-centric views over planetary ecological limits.

Conclusions:

• Acknowledging AI's materiality and ecological demands is essential for planetary health.
• Moving beyond anthropocentric narratives in AI design can help align technology with ecological limits.
• Responsible AI development is necessary to ensure technology serves science and all life on Earth.