Part 4/17:
Karen How criticizes political and corporate use of “AI,” likening it to the vague concept of transportation—ranging from bicycles to rockets. Such an analogy reveals the dangers of broad, unfocused policies, such as funding AI for healthcare or climate solutions without specifying what forms of AI are involved or suitable.
Environmental Costs: The Hidden Toll
The development and deployment of AI are energy-intensive, relying heavily on data centers that consume substantial resources. Currently, global data centers account for about 3-3.5% of CO₂ emissions—a figure that is poised to grow rapidly. Projections suggest that AI-specific data centers could require an energy increase equivalent to adding half to 1.2 times the annual UK energy consumption within five years.