Joseph Weizenbaum Publishes Computer Power and Human Reason

Weizenbaum was the legendary creator of the ELIZA chatbot, which to this day remains relevant to the discussion around trusting relationships with AI and ascribing personhood to AIs. Weizenbaum saw people fall into this trap in the 1960s of trusting a simple chatbot as if it understood them. This book demonstrates his warning that some jobs, like therapy and judging, should never go to machines, which sounds eerily familiar to contemporary debates with (hundreds of?) millions of people confiding in ChatGPT and Claude.

Since ELIZA, Weizenbaum was disturbed to see people confide in his chatbot as if it actually understood them, with some psychiatrists at the time even proposing it for therapy. The book sought to lay out why precisely we shouldn't be trusting these machines.

It distinguished deciding, which computers can be programmed to do, from choosing, which requires judgment. Weizenbaum saw where this was all headed and used his deciding-vs-choosing position to argue against giving machines certain jobs (cough, cough in the future), whatever they become capable of.

It made Weizenbaum one of the first critics of AI from within computer science. He's still celebrated as such today by artists like Eva and Franco Mattes.