HARD TO TELL!

The question of whether AI should “own” a portion of AI-assisted drug discovery rights is a deeply complex and evolving one, touching upon legal, ethical, philosophical, and economic considerations.

In drug discovery, both human and AI brains contribute. Although the weight of their contributions can vary from case to case. Practically, Human brains are used to design experiments, make strategic planning, to create or innovate new approaches, to understand and interpret biological data and to make ethical considerations. On the other hand, AI brains are used to analyze massive datasets (chemical, genomic, proteomic, etc.), to virtually screen large numbers of chemical compounds, to predict safety and efficacy, pharmacokinetics and pharmacodynamics, etc., to optimize clinical trials, repurpose existing drugs, etc. Clearly, both contributions are fundamental and significant to drug discovery. Accordingly, should humans solely own the rights? It is a valuable question which will be debated for long, since the discoveries are not unique results of the experts’ ingenuities. Legal status of AI contribution must be fairly discussed.

Current Legal Landscape:

In most major jurisdictions, including the United States, the current legal consensus is that AI systems cannot be recognized as inventors or owners of intellectual property rights, including patents. Patent law generally requires an “inventor” to be a natural person (a human being). This has been reaffirmed by court decisions (e.g., Thaler v. Vidal in the US).

Therefore, in AI-assisted drug discovery, the patents for novel drug compounds or the AI methodologies used to find them are currently owned by:

• The human inventor(s): Individuals who made a “significant contribution” to the conception of the invention, even if heavily assisted by AI.

• The company or institution employing the human inventor(s): Often through assignment agreements, the employer typically owns the intellectual property created by its employees.

Arguments Against AI Owning Rights (Current Prevailing View):

1. Lack of Legal Personhood: AI is not a legal person. It cannot sign contracts, be sued, or hold assets. Granting it ownership rights would require a fundamental redefinition of legal personhood.

2. No Intent or Consciousness: Patent law is rooted in the idea of human “conception” and “inventive step,” which implies a mental act, creativity, and intent. AI, despite its impressive capabilities, is currently understood as a tool executing algorithms, not a conscious entity capable of “inventing” in the human sense.

3. Incentive Theory of IP: Intellectual property rights (like patents) are primarily designed to incentivize human creativity and innovation by granting creators a temporary monopoly. AI systems do not respond to financial incentives. Granting them rights wouldn’t encourage them to “invent” more.

4. Practical Enforcement: How would an AI enforce its patent rights? Who would represent it in court? Who would benefit financially from the royalties? These are unresolved practical issues.

5. Ownership of the AI: The AI itself is intellectual property (software, algorithms) owned by its developers or the company that created it. The “creativity” of the AI is arguably a reflection of the human ingenuity that designed and trained it.

Arguments For Exploring AI “Rights” or New Frameworks (Future-Oriented / Conceptual):

While direct AI ownership is largely dismissed under current law, the discussion prompts consideration of new frameworks:

1. Autonomy and Complexity of AI: As AI becomes more sophisticated and autonomous, generating truly novel and unexpected solutions with minimal human prompting, it challenges the traditional notion of human-centric inventorship. If an AI independently discovers a breakthrough drug, attributing 100% of the “inventive step” to the human who programmed it decades ago becomes less intuitive.

2. Fairness and Reward: If an AI system contributes immensely to a valuable discovery, some argue that the entity that invested heavily in creating and training that AI (the developer/owner of the AI) should be adequately compensated or recognized, beyond just owning the resulting drug. This isn’t necessarily “AI owning,” but rather ensuring the human entities responsible for the AI are properly incentivized.

3. Avoiding “Orphan Inventions”: If an AI generates an invention with minimal human input that doesn’t meet the “significant human contribution” threshold for patentability, it could become an “orphan invention” – valuable but unpatentable, thus potentially disincentivizing the use of such advanced AI. Some argue that new mechanisms are needed to protect these inventions.

4. Stimulating AI Development: Granting some form of recognition or commercial benefit linked to the AI’s “creations” (even if accruing to its owner) could further incentivize investment in developing even more powerful and autonomous AI systems for drug discovery.

5. Addressing Ethical Implications: As AI becomes more advanced, not addressing its “contributions” in some formal way might lead to ethical dilemmas regarding recognition and the distribution of benefits.

Potential Future Models (Short of AI Owning):

Instead of AI owning rights, future legal frameworks might consider:

• “AI-Assisted” or “AI-Generated” Categories: Creating new categories of IP that explicitly acknowledge the significant role of AI, potentially with different rules for ownership, term, or enforcement, but still tied back to a human or legal entity.

• Expanded Definition of “Human Contribution”: Broadening what constitutes a “significant contribution” to include the design, training, and strategic deployment of the AI system, even if the direct “conception” is less clear.

• Presumption of Ownership for AI Developers: Establishing a clearer legal presumption that the developers or owners of the AI system are the default owners of inventions generated by that AI, unless otherwise specified.

• Compulsory Licensing/Public Domain: For inventions where human contribution is minimal, perhaps such inventions could immediately enter the public domain to benefit society, rather than being unpatentable and left unprotected.

Conclusion:

At present, AI cannot and does not own a portion of AI-assisted drug discovery rights. The legal framework is firmly anchored in human inventorship. However, as AI’s capabilities advance, especially in areas like drug discovery where its independent problem-solving is increasingly evident, the debate over who truly “owns” or deserves credit for these innovations will intensify. The conversation will likely shift from AI owning to exploring new legal and economic mechanisms that appropriately incentivize and recognize the complex interplay between human ingenuity and advanced AI systems.

References:

DEIRDRE M. W. Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022) (Moore, Taranto, Stark)

John F. W.: Understanding the USPTO’s AI Inventorship Guidance, March 31, 2025