Artificial intelligence (AI) is transforming diagnostics—discovering disease biomarkers, predicting treatment response, and designing novel imaging algorithms faster than ever before. Yet, as these systems begin generating inventions autonomously, one question continues to challenge patent law worldwide:
If AI creates a new diagnostic tool, can it be named as the inventor?
Welcome to the frontier of AI inventorship, a debate reshaping the foundations of intellectual property law.
1. The Rise of AI in Diagnostics: From Assistive to Autonomous
AI’s role in healthcare has evolved dramatically. Once limited to assisting clinicians in interpreting data, AI systems now:
- Discover new biomarker signatures from genomic and proteomic datasets.
- Design neural network architectures that outperform human-coded models for medical imaging.
- Generate new diagnostic algorithms that can identify diseases long before symptoms appear.
These are not mere tools—they’re creators in their own right, capable of generating new, useful, and non-obvious inventions. If the creative leap comes from the AI itself, without a human specifying the combination, who is the rightful inventor? This matters because inventorship determines ownership and commercial control. Without recognition, AI-generated inventions risk being unpatentable, pushing innovators toward trade secrecy. Yet patent law, as currently written, assumes that every invention must have a human inventor.
2. The DABUS Precedent: The Spark of the Inventorship Debate
Dr. Stephen Thaler’s AI system, DABUS (Device for the Autonomous Bootstrapping of Unified Sentience), sparked this debate by autonomously generating two inventions—a fractal food container and a neural-flashing light device. Patent filings across multiple jurisdictions yielded a global legal experiment:
| Jurisdiction | Decision | Key Reasoning |
| United States | 16/524,350 Rejected | “Inventor” under the Patent Act must be an individual, meaning a natural person (USPTO, 2020; Federal Circuit, 2022). |
| European Patent Office (EPO) | EP3564144A1 Rejected | The EPC requires the inventor to have a legal personality; AI does not. |
| United Kingdom | GB2592777A Rejected (UKSC 2023) | Patent law presupposes human inventorship; AI cannot hold rights. |
| South Africa | ZA202103242B Accepted | First country to grant a patent listing DABUS as inventor (formality-based acceptance, not a full legal recognition). |
| Australia | AU2019363177A1 Initially accepted, later overturned | The Federal Court reversed its earlier decision, aligning with other jurisdictions. |
The message from most IP offices: AI can assist, but not invent.
3. Human Proxy Inventorship & Emerging Policy Shifts
Under current law, inventorship must always trace back to a human. When AI contributes to an invention, the person who designs, trains, or directs the system is listed as the inventor. This “human proxy” model, while legally practical, raises questions about whether it truly reflects the creative source.
Governments and IP offices are actively studying how to adapt. Recent initiatives include:
- USPTO’s 2024–2025 AI Inventorship Study: Exploring whether AI contributions should confer rights or recognition, with feedback from tech and biotech industries.
- WIPO Conversation on AI and IP (Geneva): Global dialogue on harmonizing approaches to AI-generated inventions.
- European Commission Reports: Discussing potential updates to inventor attribution under the EU AI Act framework.
4. Looking Ahead: The “Machine-Human Partnership” Era
The future of diagnostics will be co-creative. AI systems won’t replace inventors — they’ll redefine what inventorship means. In the coming years, we may see:
- Recognition of AI as a co-inventor (acknowledged but without legal rights).
- New categories like “machine-assisted invention”.
- Cross-jurisdictional frameworks balancing innovation incentives and public trust.
- Integration of AI disclosure standards in patent filings (already hinted at by the EPO).
AI systems naturally evolve as they learn, adapt, or retrain on new data, raising an important question: could such changes make the AI a “second inventor”? Under prevailing patent frameworks, the answer remains unequivocally negative. Even when an AI generates new or modified outputs, inventorship remains with the humans who design, train, or direct it—keeping humans legally in the driver’s seat, even if the steering wheel is increasingly guided by code.
Final Takeaway
The DABUS filings highlight a critical lesson for innovators: when AI inventorship remains legally unsettled, misaligned filings can result in costly rejections across jurisdictions.
At MaxVal, our Search and Analytics experts help innovators avoid such pitfalls through deep patentability insights and legal-technical expertise. We ensure that every invention, whether human or AI-assisted, is positioned for success by evaluating patentability early and providing actionable search insights that help protect innovation while optimizing costs. In the evolving age of AI and diagnostics, strategic foresight—not just invention—drives lasting IP value.
