The recognition pathway: reconceiving inventorship and obviousness for iterative CRISPR-AI convergence

Clustered Regularly Interspaced Short Palindromic Repeats -Artificial Intelligence (CRISPR-AI) convergence generate therapeutic candidates via iterative human-machine workflows that disrupt the inventorship criterion of patent law based on conception. In CRISPR-AI studies, AI systems suggest thousands of guide RNAs or Cas variants; human researchers select, confirm, and optimize them via experimental feedback loops that conceal the locus of conception. The existing patent doctrine, which demands a clear and final concept in the human mind, fails to align with inventions in which AI comes up with solutions that human researchers had not anticipated. This Article advances a recognition pathway for CRISPR-AI inventorship that maintains the natural-person baseline while attributing inventorship to humans who exercise claim-defining technical judgment in identifying patentable value within AI outputs. The Article reconceives the person having ordinary skill in the art standard for AI-enhanced biotechnology, rejecting both inflexible pre-AI baselines and the 'computer having ordinary skill in the art' proposal. A five-factor test is employed to assess AI-generated CRISPR inventions on non-obviousness in terms of problem structure, predictability of solutions, and quality of human judgment. The framework provides a computational biotechnology human-accountability grammar that is administrable, but not legally granted to machines.