Integration Of Ai In Patent-Office Decision Support And Ethical Oversight.
1) AI in Patent-Office Decision Support
AI is being increasingly used in patent offices to support examiners rather than replace them. Its applications include:
Prior art searches
Classifying patent applications
Drafting office actions
Checking for patentability (novelty, inventive step, sufficiency of disclosure)
While AI can improve efficiency, it raises ethical issues:
Bias in AI tools
Lack of transparency (black-box algorithms)
Responsibility for decisions
Inventorship recognition
Patent offices must ensure that AI augments human decision-making without compromising legal standards or fairness.
2) Case Law and Policy Examples
Here are five key cases illustrating how courts and patent offices have addressed AI-related patent issues:
Case 1: Japan – DABUS AI as Inventor (Tokyo District Court)
Facts:
Dr. Stephen Thaler submitted a patent application naming his AI system, DABUS, as the inventor.
Patent Office Action:
The Japan Patent Office refused the application, stating that only natural persons can be inventors.
Court Decision:
The Tokyo District Court upheld the JPO decision. It emphasized that AI lacks legal personhood, so inventorship must belong to a human.
Significance:
Reinforces human accountability in patents.
Clarifies that AI can assist inventing but cannot legally be an inventor.
Case 2: DABUS Patent Attempts Globally
Facts:
Thaler tried to get patents granted naming DABUS as inventor in multiple countries, including the U.S., UK, EPO, New Zealand, South Korea.
Decisions:
USPTO: Refused, stating inventors must be natural persons.
EPO: Rejected applications under EPC rules.
UK High Court: Upheld rejection.
Significance:
Globally, courts are consistent: AI cannot hold inventorship.
Establishes a legal and ethical precedent for human oversight in AI-assisted inventions.
Case 3: EPO – T 0161/18 (Sufficiency in AI Inventions)
Facts:
Patent claimed a neural network invention, but the description lacked details on training data and model design.
Decision:
The EPO Board held the disclosure insufficient under Article 83 EPC. A skilled person could not reproduce the invention based on the description.
Significance:
AI inventions require detailed, transparent disclosure.
Ensures AI does not create opaque inventions that cannot be executed by humans.
Case 4: USPTO – Ex Parte Desjardins (Machine Learning Patent)
Facts:
A patent on sequential training of AI models was initially rejected under §101 as abstract.
Decision:
The USPTO Director vacated the rejection, stating the PTAB failed to properly assess patent eligibility.
Significance:
Shows internal ethical oversight within patent offices.
AI-related inventions must be evaluated carefully to avoid misclassification.
Case 5: USPTO Guidance on AI Inventorship (2025)
Policy:
AI cannot be listed as an inventor.
Only humans who conceive the invention can be inventors.
AI is a tool, not an independent inventor.
Significance:
Provides formal ethical guidance.
Ensures consistency in decision-making when AI is involved in invention.
Case 6: European Court – Inventorship vs. Ownership in AI Patents
Facts:
A European applicant claimed an AI system generated a chemical compound. The ownership of resulting IP was disputed.
Decision:
The Court ruled the patent must list a human inventor, but ownership can reside with the person or company controlling the AI.
Significance:
Separates inventorship (human requirement) from ownership (can be corporate).
Clarifies legal accountability in AI-assisted patents.
3) Ethical Issues Highlighted by These Cases
Human Inventorship: AI cannot replace human inventors.
Transparency: AI-assisted inventions must be sufficiently disclosed to allow replication.
Decision Oversight: AI tools in patent offices must be monitored to avoid errors or bias.
Accountability: Humans are ultimately responsible for decisions influenced by AI.
Fairness and Bias Prevention: Offices must prevent AI from embedding systemic bias in examination results.
4) Conclusion
AI in patent offices is a decision-support tool, not a decision-maker.
Courts and patent offices globally enforce human accountability and ethical oversight.
Key legal lessons: human inventorship, transparency, sufficiency of disclosure, and monitoring AI influence.
These cases show that AI can assist in innovation, but law and ethics maintain a human-centric framework.
RELATED Blog
comments