Should a generative-AI system be able to count as an "author," and what's the problem with how such systems are trained?

Generative AI trains on artists' works without their consent, which raises both an ethical problem (uncompensated, unconsented use) and a legal puzzle (can a non-human system be an author/rights-holder at all?).

Two intertwined issues:

• Training data. Generative models use artistic works as training material — typically without the consent of the artists whose work makes the output possible. This raises questions of fairness and transparency about when and how AI was used.
• Authorship. Can a non-human system be an author in the legal sense? Authorship law was built around human creators with intentions and rights; an AI has neither, yet it produces work (e.g. "The Next Rembrandt," an AI-generated painting in Rembrandt's style). The category strains.

The critical-thinking point: new technology doesn't just raise "does it work?" but forces us to re-examine concepts — authorship, consent, originality — we assumed were settled.