Will Fair Use Save AI from Common Sense?

This essay argues that the artificial intelligence fair use debate has become too technical and too detached from economic common sense. Modern AI systems depend on massive volumes of human-created expressive works. Those works are not incidental.

They help supply the fluency, style, cultural competence, and expressive range that make generative AI valuable. Yet AI developers seek to use those works without payment while paying for every other major input, including engineers, chips, cloud computing, energy, and capital. The central question is not whether AI is useful. It is whether fair use should allow a commercial technology to treat creative works as a free industrial input.

Even if traditional market-harm doctrine struggles to capture diffuse harm to creative labor markets, the underlying asymmetry remains: AI systems extract value from expressive works and then compete with the markets that produced them. The innovation defense is serious. Uncompensated use may sometimes accelerate general-purpose technologies and produce broad public gains.

But if fair use is being used to shift the cost of AI’s expressive inputs onto creators, that trade-off should be named clearly. Fair use may answer the legal question, but it should not obscure the common-sense one: who should bear the cost of the expressive works that make generative AI valuable?