This week, IPWatchdog Unleashed featured a detailed discussion on the intellectual property (IP) challenges facing companies in the rapidly evolving artificial intelligence (AI) sector. The conversation brought together Ed Russavage and John Strand, both shareholders at Wolf, Greenfield & Sacks, PC, who recently participated as speakers in the AI 2025 programme.
The discussion took place in front of a live studio audience as the programme drew to a close, providing an opportunity to explore AI-related IP risk from a client’s viewpoint as well as strategic approaches to mitigate these risks.
John Strand outlined the cautious stance of many companies currently navigating the AI landscape: “It’s a lot of wait-and-see unfortunately,” he noted. Strand explained that most clients seeking risk assessments are mid-sized to larger, data-driven companies aiming to leverage AI to unlock new value from their data. A central concern, he said, involves understanding the IP risks associated with using third-party AI models and platforms, particularly those based on open AI: “If I use their modelling or them as a back-end for whatever I’m going to do with my data, what’s my risk assessment coming out of that and how that’s going to happen.”
Ed Russavage, a patent attorney with three decades of experience in the software industry, emphasised the importance of a proactive patent filing strategy in this fast-developing field. “We just tell folks file early file often and try to file something good,” he explained. He noted that they frequently utilise provisional patent applications and encourage filing within weeks or months rather than years, contrasting this with traditional approaches for software companies. Russavage advised focusing on innovations with a longer lifespan, such as user interface features or application programming interfaces (APIs), which can remain relevant through multiple software iterations. He cited Microsoft’s spell-checking GUI feature as an example of a patentable innovation with enduring value.
The conversation then turned to the significance of trade secret protection for AI companies. Given that data and the insights derived from data are often the most valuable assets, the decision between patenting and trade secrets remains critical. Russavage commented, “If you can’t keep it a secret certainly it can’t be a trade secret and so that’s when we lean towards patenting or relying on some other type of protection—contracts or what-have-you but data is going to be the big valuable thing.”
He elaborated that while many companies possess impressive technology, their most valuable asset is often the data they hold. The ability to maintain large, proprietary data sets provides a competitive advantage in developing more accurate AI models. Russavage also highlighted potential future litigation concerns: “I’ve not seen this firsthand yet but I know for sure that in any trade secret litigation one of the first things will be discovery requests about what AI software’s you’re subscribing to and subscriptions out there and how you’re putting your data out there because that will be an easy argument if you’ve released it to a third party because you didn’t have the right contractual relationship with whoever you’re giving your data to for use.”
The broader IP challenges facing the AI industry were also explored, including trademark issues tied to an immature sector where definitions and generic terms are still unsettled. Additionally, potential liabilities relating to copyright infringement could arise from AI platforms using training data without permission and from the outputs generated by AI tools.
Listeners interested in the full conversation with Ed Russavage and John Strand can access the podcast via typical podcast platforms, IPWatchdog Unleashed on Buzzsprout, or the IPWatchdog YouTube channel. The discussion sheds light on the evolving legal landscape as companies strive to protect their innovations and data in the transformative AI industry.
Source: Noah Wire Services