OpenAI’s Terrifying Vision for an ”AI-First” Society

Something is deeply wrong here...

Cartoon by Tom Fish Burne

When I was a young law student, I was once approached by a homeless immigrant outside of the law faculty in Copenhagen, who begged me for a cigarette. He told me that he used to be a promising young scientist with a Ph.D. degree in physics from a fancy European university. Now, he was forced to sleep on a cardboard box and had “no pride”. Even though I don’t remember much of the exact exchange – and of course, he could be lying - the impression I walked away with was that anyone could end up in his shoes.

I am indeed worried that many young people with talent today will. A new Gallup survey shows that the percentage of respondents between age 14 and 29 who said they felt hopeful about AI had declined from 27% last year to 18% as of February and March. Also, “48 % of working Gen Zers believe the risks of AI in the workforce outweigh its benefits”. No way that a new consumer/work technology has ever generated so little excitement and so much concern among those who are supposed to be its main adopters. Is AI having a “PR crisis” or is there something else going on?

Besides AI’s potential to absorb middle class jobs, the vulnerable, global supply chain it relies upon, and the extreme amount of resources that goes into creating cutting-edge models, the competition for entry-level jobs is fierce. I know from experience that companies tend to treat candidates like a bunch of data points, while they expect perfect credentials and love letters in return.

Increasingly, bosses are convinced that AI is a worker-replacement technology, thanks to billionaire influencers like Sam Altman, whose shameless self-promotion entice investors, board rooms and executives the world over. Some people have no pride and sleep on cardboard boxes; other people have no shame and become billionaires.

AI Is Porn for Cost-Cutting Corporations

In 2019, OpenAI was a non-profit organization that refused to launch GPT-2 to the wider public due to fears about how it could be misused. Needless to say, GPT-2 is a laughably primitive model by today’s standards. How much have things changed!

In the paid section below, I will comment on OpenAI’s document, Industrial Policy for the Intelligence Age: Ideas to Keep People First.