Artificial intelligence and machine learning are revolutionizing the medical device industry, yet these advancements bring substantial patent and regulatory hurdles that corporations need to navigate carefully.
Gaining Food and Drug Administration (FDA) approval while securing intellectual property (IP) protection are two crucial but complex tasks facing companies developing AI-powered medical devices. As these technologies advance and newer guidelines emerge, IP practitioners must guide their clients through the intricacies of patent and copyright law without jeopardizing investments in AI-enabled technologies. Recent regulatory changes highlight these evolving challenges.
Among multiple considerations is the effect of using AI to create software as a medical device on the ability to obtain patent or copyright protection. Both the Copyright Office and the US Patent and Trademark Office have issued specific guidelines on when AI-enabled activities are eligible for IP protection. Companies must evaluate how these regulations impact their software designs and core functionalities.
Regulatory Considerations
Although software as a medical device has been around for decades, advances have led to new-generation devices leveraging AI models for classification, prediction, diagnosis, or treatment of various conditions. While these devices promise clinical benefits and improved patient outcomes, an improperly designed or trained AI system could result in incorrect outputs, posing significant patient safety risks. Consequently, the FDA and Congress have been implementing a new regulatory framework for these innovations.
Historically, developers used either the premarket approval process or the 510(k) process to gain regulatory approval. However, assuming an AI-enabled software obtains approval, any updates or modifications often require a new submission process, potentially causing delays in making software updates commercially available. Thus, the FDA initiated a software pre-cert pilot program to streamline lifecycle monitoring and reduce the bottlenecks associated with regulatory approvals.
IP Considerations
For each regulatory pathway, developers must provide the FDA with comprehensive information about their AI models, which raises several IP concerns. Statements made during the FDA approval process have implications for patent inventorship and copyright authorship. Furthermore, medical software often faces challenges under Section 101 of the Patent Act regarding patent eligibility.
Some developers may opt for trade secret protection over patents to secure critical aspects of their software. However, under the new FDA guidelines requiring detailed information about AI models and training data, companies must exercise caution when submitting proprietary data to ensure it remains a trade secret. The 2022 Department of Commerce guidance underscores that inconsistent statements across different regulatory bodies could violate duties of candor and fair dealing.
For developers, ensuring that representations made to both the FDA and USPTO are consistent will be essential to prevent future enforceability issues surrounding their patents.
This article is based on an analysis by WilmerHale attorneys Alexis Cohen, Gilbert Smolenski, and Xanadu Halkias. The original source article can be found here.