As artificial intelligence (AI) systems become central to business innovation, trade secret law is crucial for protecting the proprietary algorithms and datasets that drive them. However, AI presents unique challenges, requiring strategies to keep these competitive advantages secure in an increasingly digital landscape.

The protection of intellectual property has undergone a shift. For years, patents and patent litigation were the preferred way for companies to protect and enforce their valuable intellectual property rights. Recently, however, patent litigation has declined, and trade secret litigation is on the rise. This is a result of many factors, including, but not limited to, the rapid change in technology making the years-long process to obtain a patent unsuited to protect inventions that lost their cutting edge quicker, the all-too-frequent reduction of patent damages awards by trial courts after jury verdicts, and the risk of patent invalidation in inter partes proceedings at the Patent Trial and Appeal Board (PTAB). Trade secret litigation does not have these problems.

Adding to the ledger in favor of trade secret protection over patent protection is the greater ability to protect AI related information. Generally speaking, a trade secret protects: i) information; ii) that has economic value from not being generally known by others; and iii) is secret. Trade secrets are often a formula, process, device, or compilation regularly used in business. Factors considered by courts in determining if something is a trade secret include: i) the value of the information to a company and its competitors; ii) the extent of knowledge of the information to those within a company and the actions taken to guard and limit its dissemination; iii) the amount of money spent in developing the information; and iv) the difficulty for others to obtain or create the information. A trade secret is entitled to protection for as long as it is secret.

Patent Challenges for AI

Inventions created with AI assistance are difficult to protect under the U.S. patent system. AI systems cannot be listed as sole inventors under the Federal Circuit’s 2022 Thaler v. Vidal opinion, which requires inventors and co-inventors to be natural persons. While inventors are not prohibited from using AI tools to assist in the ultimate inventions, the natural person inventor must still be the initial source of the conception of the patented idea and provide substantial efforts. Thus, one can see a future in which patent invalidation efforts investigate the role of AI systems in the conception of the patented idea. The inventor will be required to prove his or her inventive steps through testimony to ensure an AI system was not an inventor.

Trade Secrets are Flexible

AI-created inventions have no such restrictions if they are protected as trade secrets. The AI invention just needs to be secret and have that secrecy maintained however it is monetized. Nothing in the definition of a trade secret prevents the protection of information generated by AI or the AI system itself. In fact, trade secret protection is well suited to protect AI-generated inventions and AI systems themselves. These systems are based on proprietary algorithms and curated data (information) that are usually used to aid or replace a mundane human task (economic value) where the algorithm and data set are not presented to the user (secrecy). Developing AI tools and systems that are adopted by a critical mass in general or in a specific niche can result in an astronomical value to the developing company. Charging $20.00 or less per month for use (a sum supported by the general-public market to say nothing of an industry-tailored system where monthly charges could be orders of magnitude greater) quickly yields eight figures or greater monthly revenue. Yet, companies can also take advantage of AI tools and systems internally without creating an external product. Certain companies will find value in developing their own systems or modifying a base open-source system for their own internal needs. Such an AI system may not be sold as a product but could help make a company’s internal operations more efficient or powerful giving it an economic advantage over its competition. These examples demonstrate the flexibility of trade secrets to protect different types of AI tools and systems.

Key Steps for Protecting AI as Trade Secrets

The AI tools and systems that are developed by companies are themselves valuable trade secrets. However, because of the inherent nature of these tools and systems, their protection requires special strategies. Initially, companies must have a robust trade secret protection scheme generally. This includes the use of restrictive covenants preventing the disclosure of the specific AI tool and system information through employment contracts, the use of additional Non-Disclosure Agreements for specific projects, and designating algorithms and data sets as “confidential.”

Next, because of the nature of AI tools or systems, additional protections should be implemented to prevent their unauthorized disclosure. The underlying algorithm and datasets should be restricted by password protection. Access should be granted only to those employees necessary to modify the underlying system. For extra security, modification access should be restricted to dedicated terminals not connected to the internet and a restricted internal network, if any, to the extent this is practicable. No external drives should be allowed to access the modification terminal. Companies should also ensure the algorithm and datasets cannot be copied and consider barring cell phones or other recording devices from access points to prevent screenshots of trade secret information.

Now, Keep Your Secret

These protections may seem draconian, but companies need to have systems in place to protect them from a disgruntled employee intent on taking harmful action for a perceived (or real) slight. Trade secrets only have value if they are secret. If an AI algorithm or proprietary data set is publicly disclosed, the value is lost. If it stays secret, the company has a valuable asset that maintains its value in perpetuity – or until the next generation is developed.

Image Source: Deposit Photos
Image ID:5850440
Copyright:almagami