California's AB 412 Still Demands AI Developers Do The Impossible

As I sit at my desk, surrounded by the hum of computers and the soft glow of screens, I'm reminded of the Herculean task that California lawmakers have set for AI developers. The latest iteration of A.B. 412, a bill that demands AI developers to identify and disclose copyrighted works used to train generative AI systems, is a perfect example of well-intentioned legislation gone awry. The problem is, it's practically impossible to comply with this law.

A.B. 412 sounds simple: just have AI developers create and keep a list of all the registered copyrighted works they use in AI training. That may seem straightforward, but in practice, it's anything but. The U.S. Copyright Office, the supposed authority on all things copyright, doesn't even have a machine-readable list of copyrighted works. And many copyright holders can get a copyright without even depositing a publicly viewable sample of the work. For instance, software companies may register copyright on proprietary code without revealing it to the public. This makes it nearly impossible for developers to cross-reference online data against a copyright system that simply wasn't designed to do so.

The internet, that vast and wondrous tapestry of human creativity, is a messy place. Copyright information is often incomplete, unavailable, or impossible to verify. One image may be registered with the copyright office, while the next is licensed under a free Creative Commons license, and the next is public domain. A message forum user might post an original story, photograph, or poem without any indication of ownership or registration status. It's like trying to find a needle in a haystack, except the haystack is on fire and the needle is a tiny, invisible speck.

Supporters of the bill often frame it as a way to help creative workers have some leverage against Big Tech, but the bill reaches much further than the big AI companies. Its definition of "developer" extends to anyone who makes a generative AI model available to Californians. That includes indie developers tinkering with an existing model, open-source initiatives, nonprofits, and other non-commercial efforts. Recent amendments added exemptions for universities and government entities, which is important, but that still leaves out a vast swathe of non-commercial tech work that's done by people without full-time jobs in government or academia. Large companies will hire compliance teams and lawyers to navigate this complex landscape, but for small-time developers, it's a recipe for disaster.

If California passes A.B. 412, its impact will go far beyond the large AI companies we read about in the headlines. It will stifle innovation, create a barrier to entry for new developers, and effectively lock in the power of the largest companies in AI. The Electronic Frontier Foundation (EFF) has submitted an opposition letter to the California Senate Privacy Committee, explaining why they believe A.B. 412 is simply unworkable. As a writer, I've seen firsthand the importance of access to information and the freedom to create. California's A.B. 412 is a step in the wrong direction, and I fear it will have far-reaching consequences for the tech industry as a whole.

As AI continues to evolve and become an increasingly important part of our lives, it's essential that lawmakers take a step back and reevaluate the impact of their legislation. A.B. 412 may have been well-intentioned, but it's a solution in search of a problem. By recognizing the complexity of the issue and the impossibility of compliance, California lawmakers can work towards a more nuanced and effective solution. The future of AI development depends on it.

• A.B. 412 is a bill that requires AI developers to identify and disclose copyrighted works used to train generative AI systems.
• The bill is currently being considered by the California Senate Privacy Committee.
• The Electronic Frontier Foundation (EFF) has submitted an opposition letter to the committee, explaining why they believe the bill is unworkable.
• The bill's definition of "developer" extends to anyone who makes a generative AI model available to Californians, including indie developers, open-source initiatives, and nonprofits.
• Recent amendments added exemptions for universities and government entities, but still leave out non-commercial tech work done by people without full-time jobs in government or academia.