Generative AI: Use With Caution To Avoid Risking Trade Secrets
Artificial intelligence (AI) technology is changing every industry, but it could also change the status of a company’s trade secret information. If an employee enters company trade secret information into a generative AI application, the trade secret status could be destroyed, depending on the application’s governing terms of use.
Under many terms of use, the application itself can potentially reuse information entered by a user (“input”) by further training its algorithms with inputs, presenting a risk that an input could be incorporated into a response to a third party, possibly revealing sensitive information. And once an input is entered into a generative AI application, it cannot be retrieved. Another risk is that the company supporting the application could potentially review, release or sell the information provided to third parties unless the license agreement expressly prohibits doing so.
These risks are elevated by the fact that studies show a majority of employees who use generative AI tools like ChatGPT have not informed their employer that they are using them. It would appear that AI use policies have not kept pace with the growth of AI, specifically generative AI.
Trade secret policies should be modified to specifically direct (or prohibit) the use of generative AI, with specific guidelines on using company trade secret information as an input. Some companies may also need to revisit employment agreement provisions regarding the use of confidential information. An outright ban on the use of generative AI may well put a company at a competitive disadvantage. However, companies should be thoughtful about which employees are permitted to use AI and generative AI for work, how they are permitted to use it, and what information they are permitted to provide. Moving beyond policies, companies should also consider using software applications that can prevent certain key words or phrases from being entered into a generative AI application. Finally, companies should also thoroughly review and evaluate the terms of their agreements governing the use of generative AI applications to ensure that they have adequate protection against company information being disclosed to or used by third parties.
New Circuit Split Raises Uncertainty About the Recovery of Unjust Enrichment Damages
Federal courts have traditionally held that unjust enrichment damages are available in trade secret cases for costs avoided by a defendant company in developing its own trade secrets, even when the trade secret holder has not suffered corresponding harm beyond lost profits. The U.S. Court of Appeals for the Second Circuit recently issued an opinion that takes a different stance.
The plaintiff company, a trade secret holder that developed software used by healthcare insurance companies, established that a former subcontractor misappropriated its trade secret information and was awarded damages for both lost profits and developmental costs that were avoided by the subcontractor.
On appeal, the Second Circuit recognized that in some circumstances, a trade secret plaintiff can recover both actual losses and unjust enrichment if those damages are not duplicative. The court concluded that unjust enrichment damages are meant to make the trade secret holder whole when damages for actual losses are not sufficient, such as when the value of the trade secret has been compromised, or when the defendant has somehow benefitted from the use of the trade secret in a way that is not compensable as lost profits.
In this case, the Second Circuit found the trade secrets had not lost value, noting that the plaintiff had obtained an injunction that ended the defendant’s use of the trade secrets and its ability to profit from avoiding costs to develop its own trade secrets. Therefore, the court determined that the lost profits award was sufficient and struck the award of unjust enrichment damages for avoided costs as duplicative.
This Second Circuit opinion is at odds with decisions from the Third and Seventh Circuits, where the courts affirmed awards of unjust enrichment damages for avoided costs beyond damages for lost profits to account for the “head start” the misappropriating parties gained in the marketplace through their misconduct. This circuit split will likely form the basis of new challenges to claims of unjust enrichment with no ultimate resolution in sight, as the U.S. Supreme Court has denied cert.
Ohio Appeals Court Refuses To Modify Noncompete Agreement, Instead Finds It Unenforceable
In February, an Ohio court of appeals issued an opinion upholding a lower court’s decision not to enforce a noncompete agreement against a former Kross employee because it was overly broad in geographic and temporal scope. Kross v. Acquisition Co. v. Groundworks, 2024-Ohio-592 (Ct. App.). The noncompete restricted a former sales representative from working in the specified field for two years and throughout Ohio and Kentucky. Not only did the court find the noncompete unenforceable, it refused to modify the agreement.
Typically, even when Ohio courts find restrictive covenants unreasonable as written, they modify the provisions to make them more reasonable and enforceable by narrowing the restrictions. This is known as “blue-penciling.” The Kross court found that Raimonde v. Van Blerah, which set state precedent for evaluating noncompete agreements, permitted blue-penciling, but did not require it. Thus, the appeals court found that the lower court was justified in refusing to modify the noncompete agreement at issue.
Given the Kross decision, employers are cautioned against relying on blue-penciling to recraft restrictive covenants through litigation. This case is also yet another call for employers to revisit noncompete and nonsolicitation restrictions as a precautionary measure to ensure that they are only including restrictions that are necessary to protect a company’s legitimate business interests. Otherwise, employers risk courts deeming noncompete agreements wholly unenforceable.
This newsletter may be reproduced, in whole or in part, with the prior permission of Thompson Hine LLP and acknowledgement of its source and copyright. This publication is intended to inform clients about legal matters of current interest. It is not intended as legal advice. Readers should not act upon the information contained in it without professional counsel.
This document may be considered attorney advertising in some jurisdictions.
© 2024 THOMPSON HINE LLP. ALL RIGHTS RESERVED.
