Can AI's data, ethics, and trust problem be solved by class actions or rearguard ethical campaigns?

San Francisco, CA-headquartered Anthropic has long claimed to be a "high trust, low ego" company – the ethical choice, no less. But its $1.5 billion settlement of the copyright class action brought against it by US authors demonstrates the absurdity and hypocrisy of that claim.

Anthropic's Claude AI was trained, in part, on seven million books scraped from pirate sources such as Library Genesis (LibGen) and the Pirate Library Mirror – on the orders of a CEO, Dario Amodei, who is known for making sweeping claims about AI's decimation of the market for entry-level jobs – and perhaps all jobs.

As with so many AI CEOs, the ego appears to have landed.

The $1.5 billion settlement covers roughly 500,000 of those books, essentially offering their authors $3,000 per title – good news, perhaps. But if a writer has spent years crafting an original work with their talent, skill ...