Karen Hao – AI Whistleblower (2)

1. Environmental and Technical inefficiency

• Unnecessary Scale – Hao contends that the massive scale of data centres being built is technically unnecessary, and that the narrative of larger being better is a sales pitch that harms the environment.
• Resource consumption – These empires are draining water for cooling and relying on fossil fuels, creating a high environmental cost that is not properly factored into their, or the publics calculations.

2. Call for Accountability and smaller AI models.

• The Bicycle approach – Instead of chasing generalized all purpose everything machines, Hao advocates for developing smaller, more specialized AI bicycles that can solve specific, high leverage problems…medical diagnostics, climate problems.
• Need for Action – Hao emphasizes that the current trajectory is not inevitable and calls for increased, democratic contestation and empathy-driven journalism to hold these companies to account.

3. The 30% Rule:

• Hao advocates for the 30% AI Rule, which creates a working framework for AI development in that, no more than 30% of the work should come directly from AI tools to maintain human agency, creativity and accountability.