Generative AI Tools: Confidentiality & IP Considerations for Every Business & Organization

As Seen in the OCBJ’s Oct. 6, 2025 Issue

Commercial Bank of California is a full-service bank and diversified financial services company serving businesses, professionals, and communities in the greater Los Angeles and San Francisco Bay areas of California. Recognized as a BauerFinancial, Inc. “Four-Star Excellent Bank” for its financial strength and stability, CBC provides the financial expertise of a major bank while maintaining a commitment to personalized service for every CBC client. More information about CBC’s custom solutions for your business is available at www.cbcal.com. CBC’s General Counsel and EVP Corporate Development Raj Shukla has overseen the company’s legal issues for several years and stays current on technology updates in the banking sector. One area of interest for many companies is the review and adoption of generative AI tools, and the issues that arise with use of these tools. This article highlights confidentiality and intellectual property issues to consider with evaluating the use of generative AI. To start with the basics, generative AI uses artificial intelligence to create content based upon an existing set of data. Many of the widely known generative AI tools are trained on large collections of data, and these data collections are typically processed and filtered before they are used. There are many publicly available generative AI tools that generate content based upon the collection of data on which they were trained. One well-known publicly available example is ChatGPT, a generative AI tool developed by OpenAI that receives prompts or queries from a user and provides an output in a conversational-style format. Generative AI tools, including an upgraded version of ChatGPT, can also be accessed through paid license fees. These tools can further be structured to generate content based upon a limited dataset tailored to a specific company or use case.

Confidentiality and Reliability Considerations with Generative AI Tools

As companies and their employees explore and use generative AI tools to create content, there are some considerations regarding confidentiality and reliability to keep in mind:

• If employees are using generative AI tools, does the company place any restrictions on how these tools are accessed? • Are the tools publicly available, and if so, how current is the collection of data that is being used to generate content from those tools?
• If the tool is being accessed through a private license, what data sources are the generative AI tool drawing from in order to generate new content?
• If the generative AI tool is creating content based on a specific dataset of the company’s information, is there any sensitive or private company information included in that dataset?
• What kind of information is being input into the generative AI tool when a user is running queries?
• Is there a process in place for evaluating and verifying the accuracy of the generated content and the sources used?

Protecting Innovation – Patents

From an IP perspective, it’s also important to understand the extent generative AI is used in any innovations that the company is seeking to protect. U.S. patent law requires that patent applicants identify the individuals who contributed to the conception and reduction to practice of the claimed inventions, and a patent can be challenged if the inventorship is incorrect. U.S. patent law also requires that the subject matter of an invention must be “eligible” for patent protection. When evaluating inventions created with the assistance of AI, the US Patent Office has issued guidance explaining that patent protection can be pursued for AI-assisted inventions when individuals make a significant contribution to the invention. Examples of the types of inventions that are eligible for patent protection include the use of an artificial neural network to detect anomalies, AI-based methods of analyzing speech signals, and an AI model that can assist with personalizing medical treatment to an individual patent.

Protecting Brands – Trademarks

When seeking trademark protection, trademark applicants must provide evidence that their marks (e.g., brands and logos) are used in commerce, but the Trademark Office has cautioned against the use of AI-generated evidence for this purpose. There has also been an increased use of generative AI tools to create marketing and advertising content, because of the speed and cost advantages that can be realized with the use of generative AI. If the source material for AI training includes others’ marks, this can increase risk that a company may be using others’ marks without authorization.

Protecting Expressive Works – Copyrights

Copyright protection is another evolving area for generative AI. The Copyright Office continues to evaluate policy issues around the scope of copyrights in AI-generated works (such as how much human involvement is required or how to attribute authorship), and various courts are considering the appropriate use of copyrighted works in the training of AI models. Use of AI-generated images can create copyright risk and potential concerns over rights of publicity when authors of AI training source material assert their rights. As companies adopt generative AI tools, keeping these considerations in mind will help to protect the company’s confidential information, minimize risk, and optimize the company’s ability to protect its innovations and brands.

Authors:

Raj Shukla is General Counsel and EVP Corporate Development of Commercial Bank of California, responsible for all legal matters as well as corporate  M&A. Previously, he held lending and credit focused roles at both publicly traded and privately held financial institutions, as well as an associate role at a law firm. He started his career in high-tech in the San Francisco Bay Area, and earned an MBA and an MRED in addition to his JD after relocating back to Southern California.

Sheila Swaroop is a partner in the Orange County office of Knobbe Martens, which she chairs the Litigation Group. She is a first-chair trial lawyer who has litigated numerous intellectual property cases over the course of her career and enjoys working with clients to protect and defend their innovations in the market. She can be reached at (949) 721-7646 or
Sheila.Swaroop@knobbe.com.