In a recent post (here), I wrote about a lawsuit that had just been filed against Microsoft’s board, alleging that the company’s directors had violated their fiduciary duties by knowingly allowing its AI development efforts to engage in copyright infringement. The case, I said, represented an example of “silent AI” – that is, the seepage of AI-related matters into various insurance coverages that were not consciously intended to provide coverage for certain exposures. The case showed how a matter that would not typically be covered under a D&O policy (copyright infringement) can translate into a potentially covered matter (a breach of fiduciary duty lawsuit).
In the latest example of this kind of lawsuit, a plaintiff shareholder has filed a derivative lawsuit against the board of Nvidia, alleging that its directors knowingly permitted its AI models to violate copyright holders’ rights and allowed violations of the Illinois Biometric Information Privacy Act (BIPA). Nvidia, the complaint alleges, has been the target of numerous copyright infringement actions, as well as class actions brought for alleged BIPA violations concerning individual voiceprints. The derivative lawsuit seeks to hold the company’s directors liable for the company’s “potentially massive liability and related costs and reputational damages” that the company faces in the underlying litigation. As discussed below, the new Nvidia derivative lawsuit, which may be found here, represents yet another example of “silent AI” in operation in the D&O context.
Background
Nvidia is best known as the manufacturer of computer-graphics hardware, particularly for its Graphics Processing Units (GPUs), but it has recently expanded into software and hardware for training and operating Artificial Intelligence (AI) software programs. Nvidia’s hardware and software are used by all “Frontier AI” companies that develop the most advanced AI systems. Nvidia AI models include multiple AI software programs called Large Language Models (LLMs).
The Lawsuit
The derivative complaint alleges that the LLMs were trained using copyrighted materials pirated from many sources. Among other things, the source materials allegedly included thousands of hours of human speech recordings. The human voice recordings allegedly were used in violation of BIPA (among other things, by failing to identify the speakers and to provide written notice or obtain a release).
The complaint alleges that the company’s unauthorized use of copyrighted or protected materials “resulted in multiple copyright holders filing lawsuits against Nvidia based on Nvidia’s failure to compensate them for downloading, copying, storing, or using their copyrighted words … for which Nvidia is now facing potential massive liability.”
The complaint alleges that the defendant directors knew of these issues with Nvidia’s use of copyrighted works without the copyrighted material owners’ permission. The complaint further alleges that the defendant directors were “well aware” of the company’s potential liability for copyright infringement, as well as potential liability for BIPA violations. Both the alleged copyright and BIPA violations have resulted in massive litigation against the company.
The complaint alleges that Nvidia is “now saddled with having to defend itself” in the litigation and is “facing potentially massive liability and related costs and reputational damages.” The complaint seeks to hold the defendant directors liable for the harm to the company caused by their alleged breaches of their fiduciary duties.
Discussion
This new derivative lawsuit against Nvidia’s board is not the first lawsuit of its type in which a company’s directors were hit with a D&O lawsuit in which it is alleged that the directors violated their fiduciary duties by knowingly allowing the company’s efforts to develop AI models and products to violate copyright holders’ intellectual property rights. As noted at the top of this post, in June 2026, a plaintiff shareholder filed a lawsuit with similar allegations against the board of Microsoft (discussed here). And as discussed here, in April 2026, a plaintiff shareholder filed a similar derivative suit involving similar allegations against the board of Adobe. (Just an aside, this new lawsuit may be the first one to raise similar-type allegations with respect to alleged BIPA violations.)
As I noted with respect to the prior lawsuits, this new lawsuit may represent something of an emerging trend, in which matters that would not otherwise be covered under a D&O insurance policy (in this case, alleged copyright infringement) are tranformed into a claim presumptively covered by the D&O policy, through the means of a follow-on breach of fiduciary duty lawsuit.
There is nothing necessarily new about the phenomenon of a follow-on derivative lawsuit. For example, we recently discussed the follow-on lawsuit trend with respect to the derivative lawsuit filed against Uber’s board, I which the plaintiff shareholder alleged that the defendant directors breached their fiduciary duties by allowing the company to pursue or continue practices and policies that allowed the company to be sued in extensive underlying sexual harassment and assault litigation. The Uber post cited numerous other prior examples in which follow-on suits claimed defendant directors had breached their fiduciary duties by knowingly allowing the company to take actions that resulted in underlying litigation against the company.
These kinds of follow-on lawsuits arguably represent examples of the oft-stated principle in the D&O arena that sooner or later everything becomes a D&O claim.
The common thread among these lawsuits is that the company has taken actions that resulted in underlying litigation against the company (usually litigation of a type that would not be covered by a D&O insurance policy) and that the underlying problem and ensuing litigation were the board’s fault (allegations of a type that presumptively are covered under a D&O insurance policy).
As we noted in connection with the Microsoft lawsuit, these follow-on derivative suits relating to underlying copyright infringement illustrate another phenomenon – that is, “silent AI,” the way in which non-covered underlying AI-related misconduct can seep into the D&O insurance policy through the follow-on fiduciary duty lawsuit. D&O insurers may well feel aggrieved because they never intended the D&O policy to pick up AI-related intellectual property liability.
These developments may be among the many reasons why the possibility of D&O insurers seeking to insert AI-related exclusions into their policies is currently under discussion in the D&O arena. With the D&O insurance market in an extended soft market phase, the possibility of the inclusion of these exclusions in most cases is a remote possibility. However, the insurance market is cyclical, and when the market eventually moves to the next phase, insurers may well seek to restrict their policies’ exposure to “silent AI.”
For now, it seems likely that as prospective litigants identify ways in which they have been harmed by artificial intelligence, the litigants’ claims may at least potentially trigger parallel follow-on lawsuits, in which the underlying allegations are relied upon to support D&O liability claims. My guess is that there will be more lawsuits in the months ahead involving allegations similar to those alleged against the Nvidia board in this new lawsuit.