Artificial intelligence is reshaping intellectual property law in patenting and trade secrets, exposing gaps across jurisdictions and adding pressure on innovation policy, according to discussions at an international symposium held in Yokohama, Japan.
Artificial intelligence is reshaping intellectual property law in patenting and trade secrets, exposing gaps across jurisdictions and adding pressure on innovation policy, according to discussions at an international symposium.*Cross-border collaboration in the life sciences is also testing legal frameworks, as disputes over ownership, disclosure and enforcement take on new complexity.
The symposium is taking place under Chatham House Rules, which don't permit identification of the individuals making the remarks at the event.
The panels highlighted a common challenge — AI and globalization are straining the boundaries of IP law and exposing fault lines in enforcement. Patentability, secrecy and ownership all face uncertainty as technology advances and partnerships expand across borders.
— AI and inventive step in pharma patents —
Panelists highlighted that generative AI is already improving molecular discovery in pharmaceutical and biotech research, and will soon likely be forcing patent offices and courts to confront disclosure, inventive step and inventorship. While some jurisdictions place no explicit requirement on applicants to disclose the use of AI, US guidance issued in April 2024 clarified that information material to patentability must be revealed. How disclosure is treated may hinge on whether AI output is characterized as a “working example” or a “prophetic example.”
Disclosure is complicated by the tension between proprietary R&D and legal obligations. Companies are often reluctant to disclose AI use that could expose sensitive processes, yet disclosure may strengthen inventive step arguments when AI-generated data supports efficacy. The consensus was that what must be disclosed is what is necessary to enable the invention, though requirements vary by jurisdiction.
The rise of AI also challenges the definition of the “skilled person.” If AI tools become part of the ordinary toolkit, what was once inventive may be deemed obvious, potentially shifting patentability thresholds.
Meanwhile, inventorship questions remain unsettled. Most courts worldwide have consistently rejected AI as an inventor, but its growing role in collaborative discovery underscores the difficulty of attributing credit solely to human inventors.
— Enforcement gaps, AI pressure in trade secrets —
In the trade secrets sphere, panelists pointed to familiar disputes over whether companies had taken “reasonable measures” to protect information, with AI introducing fresh uncertainty. Internal regulations, employee training, penalties, security protocols and third-party controls were all cited as critical, although the weight attached to each factor differs across legal systems.
The policy debate over whether trade secrets promote or hinder innovation remains unresolved, and political tensions continue to influence enforcement in some markets. Courts commonly scrutinize whether information is sufficiently particularized and whether secrecy was genuinely maintained — questions that can be especially fraught in judicial systems with broad discovery such as the US. Jurisdictional reach and the availability of remedies such as cross-border injunctions add another layer of complexity.
AI compounds these challenges. Open versus closed AI systems raise questions about whether information used in training retains its protected status. Litigation can risk inadvertent disclosure, and the extraction of data through prompts or careless employee use of generative tools has already become a live concern. Remote working practices and the use of public spaces were also flagged as weak points for secrecy management.
— IP ownership in cross-border collaboration —
Panelists examining collaboration in biopharma stressed the difficulty of dividing ownership and control of patents across institutions and jurisdictions. While rights are often split geographically or by field of expertise, cultural and institutional differences sharpen the conflict. Academic partners tend to seek joint ownership and prioritize publication, while companies focus on commercialization and guard IP tightly.
These tensions are magnified in cross-border settings. Companies rooted in consensus-driven business cultures may avoid litigation, while others use legal action as leverage to force settlements. Filing rules, data localization laws and government oversight add additional layers of risk in certain markets, particularly where national universities are involved.
Panelists noted that when disputes arise, they can quickly escalate and become emotional. Participants emphasized the importance of firewall policies, clear documentation of contributions and realistic expectations.
Practical advice included doing due diligence on potential partners, anticipating conflicts while relationships remain strong, and relying on local expertise to navigate differing legal regimes.
*AIPPI World Congress, Yokohama, Sept. 13-16, 2025
Please email editors@mlex.com to contact the editorial staff regarding this story, or to submit the names of lawyers and advisers.