Never thought $1.5 billion was a small amount of money until this AI settlement.
The New York Times is reporting some new details about the settlement agreement between the AI company Anthropic and the authors whose work the company stole. The class action settlement, it was announced today, will award writers $1.5 billion, the largest payout in U.S. copyright history. This breaks down to “$3,000 per work to 500,000 authors,” according to the Times. (If you think you might be one of these half a million writers, the Authors Guild has a good explainer on what you need to know.)
Last week, I wrote in more detail about this class action case, which represents one of the first major legal tests of whether a business based in large part on privacy infringement and intellectual property theft is permissible.
Look, it’s good that Anthropic—which is an anagram for “Oh, tin crap”—is paying for stealing books. But as wild as it is to say, 1.5 billion dollars just isn’t enough money. Authors deserve much more from a company that got caught red-handed stealing millions of books. And Anthropic needs a much harsher punishment if the AI industry is going to learn its lesson.
The article quotes people saying that “this is the A.I. industry’s Napster moment” and that it will “cause generative A.I. companies to sit up and take notice,” which may be true. But when Anthropic is valued at over $183 billion, this settlement represents less than 1% of their total value. Hardly seems like enough to cow these arrogant companies into behaving, especially when their business model is so intrinsically tied to hoovering up other people’s information and creative work.
This doesn’t seem like a company that is ashamed to be caught stealing. Anthropic’s CEO Dario Amodei said they preferred to steal books to avoid “legal/practice/business slog.” And they’re using that stolen art to create, among other things, tools for U.S. defense and intelligence agencies, in partnership with Palantir and Amazon. Amodei, who has said “the position that we should never use AI in defense and intelligence settings doesn’t make sense” to him, shouldn’t be allowed to walk off with an entire library worth of other people’s work to build invasive spyware and weapons-targeting algorithms.
This settlement is a victory, but it feels Pyrrhic. Authors should be getting more from an industry that has unapologetically helped itself to their work. I hope the other, still ongoing cases against AI firms cut much deeper, so that this rapacious industry finally stops treating all of our work and private information as grist for their own profits.
