FT column warns AI public dividend plans could shift costs abroad

In short: A Financial Times columnist argues that proposals to share AI profits with Americans could still push the physical burden of AI to other countries.

What's going on

Evgeny Morozov, writing in the Financial Times, criticizes ideas for an “AI sovereign wealth fund”, which is a state-run investment pot that would hold shares in major AI companies and pay citizens a dividend (like a national savings account that sends out checks).

The column points to two recent proposals. OpenAI chief Sam Altman suggested in July that leading US AI labs give 5 percent of their equity to a public fund. Senator Bernie Sanders has introduced a bill that would create a national fund with 50 percent ownership in big AI companies, with citizens entitled to part of the profits.

Morozov argues these plans can work even if the actual data centres move abroad. Data centres are large buildings full of computers that train and run AI, and they need lots of electricity and water (like a giant factory for computing). He says most of the heavy work of training AI can happen wherever power is cheap, while only the parts that answer users quickly need to be near users.

The column also cites examples of pushback in potential host countries. It mentions a reported $1 billion data centre plan in Kenya that stalled, and comments from Mexico’s president that her country does not need to become the “data centre of the world.”

What to watch

If governments tie public payouts to AI profits, Morozov argues there may be stronger political pressure to secure land, power, and chip access overseas. Watch for more disputes over where data centres are built, who controls them under local law, and how US chip restrictions shape competition.

Source: Financial Times