Why Trade Secrets May Be The Most Important Form Of IP For AI Innovation
Written by Ben Esplin
AI has and will become the cornerstone of transformative solutions in diverse forms of knowledge work and daily life. As technology companies race to develop these AI-driven products and services, the question of how best to protect the underlying innovations becomes paramount. While a comprehensive intellectual property (IP) strategy will almost always incorporate a mix of protection forms-patents, copyrights, and trademarks- trade secrets provide the most robust and enduring protection for the most valuable aspects of AI-driven solutions of the present day.
The unique nature of AI innovations makes them particularly well-suited to trade secret protection. The most valuable elements, including for example, proprietary training methodologies, novel fine-tuning techniques, curated datasets, development tools, and integration workflows, are typically not visible to end users or competitors. To customers, these solutions are the “black box” components, the hidden engines that drive performance and differentiation in the marketplace. Trade secret law is specifically designed to protect such confidential business information, provided that reasonable measures are taken to maintain secrecy. Unlike patents, which require public disclosure of the invention in exchange for a limited period of exclusivity, trade secrets can, in theory, last indefinitely. As long as the information remains confidential and continues to provide economic value by virtue of its secrecy, the protection endures. This is a significant advantage in the AI sector, where the pace of innovation is so rapid that the 20-year term of a patent may not align with the commercial life cycle of a given technology.
Patents, of course, have their place in an IP portfolio. They are invaluable for protecting inventions that are readily discernible from a product or service, or where reverse engineering is a legitimate risk. However, the patent system is not without its drawbacks for AI-powered solutions. For many AI-powered innovations, especially those deployed as cloud services or embedded deep within enterprise systems, the risk of reverse engineering is minimal. In such cases, the trade-off of disclosure for limited exclusivity is often not justified. Moreover, the current legal landscape for software and algorithm patents is fraught with uncertainty as we have written about before.
Given these realities, the prudent course for most technology companies is to at least consider whether the core innovations underlying their AI-powered solutions should be treated as trade secrets. We have also written before about some basic considerations and strategy around identifying and protecting trade secrets. Please contact us if you are interested in a free consultation for your company.