From ChatGPT to DeepSeek: Challenges and Implications of Large Models to the Legitimacy Theory of Copyright

With the technological upgrade of foundation models from ChatGPT to DeepSeek, under the impact of the surging wave of generative artificial intelligence, the innovation methods, contents and even the innovation subjects in the field of copyright have all been affected to a certain extent. The justification of copyright is a reflection on the premise of rights, aiming to prove the theoretical basis of the design of the copyright system. Constrained by the inherent flaws and inadaptability of the theory, the traditional theory of copyright legitimacy appears inadequate when applied to the argumentation of copyright protection for artificial intelligence generated content. Whether it is Locke’s “Theory of Labor”, Hegel’s “Theory of Personality”, or Bentham’s “Theory of Motivation” based on Bentham’s utilitarian thought, there exist insurmountable realistic obstacles. To better achieve the legislative purpose of the Copyright Law, not adopting copyright protection for the artificial intelligence generated content but instead placing it in the public domain may become a solution to the predicament of the copyright system in the era of generative artificial intelligence.