Patent Lawyers Are Shields Protecting Inventions From AI Misuse

The practice of patent prosecution is constantly evolving, driven by emerging technologies, shifting legal frameworks, and changing client priorities. For intellectual property attorneys, no recent change has the potential to impact patent prosecution more than the rapid adoption of artificial intelligence.

At the invention disclosure stage, your clients’ use of AI tools may be setting up pitfalls, particularly with inventorship and public disclosure. To secure patents in this new landscape effectively, practitioners should educate clients early on the risks of AI and establish best practices for drafting disclosures.

Navigating AI-Assisted Disclosure

The first step in the patent drafting process—the invention disclosure—traditionally ranged from napkin sketches to complex academic papers. Today, you’re more likely to receive disclosures generated with the help of AI. For inventors, AI tools make drafting faster and produce outputs that look lengthy, impressive, and superficially ready for filing.

As an IP attorney, you need to recognize and manage the misconceptions this creates. AI use often magnifies the flawed belief that long, dense, complex disclosures structured like formal patent applications are better than straightforward summaries.

Your role is to guide clients back to the value of succinctness. Dense, AI-generated text can obscure the actual inventive concepts, making it harder for you to evaluate, abstract, craft narratives, and identify worthy claim targets. Early in the drafting stage, it’s often more efficient to align with the inventor on core concepts than to parse through volumes of extrapolated AI text.

READ MORE: IP vs. AI: A New Frontier of Legal Battles to Protect Creativity

Pitfalls to Avoid

When advising clients on AI-assisted invention disclosures, you must help them navigate two critical legal hazards: challenges to inventorship and unintended public disclosure.

The Inventorship Pitfall: While the ultimate boundaries of AI inventorship have yet to be fully litigated, the US Patent and Trademark Office relies on pre-AI law to maintain that AI can’t be listed as a joint inventor. The PTO presumes that the named human inventors are correct and has declined to establish a “heightened standard” for AI-assisted inventions.

However, the pitfall lies in future uncertainty. If the PTO reacts to a potential influx of AI-assisted filings with new, more onerous rules, establishing inventorship could turn into a highly factual inquiry requiring historical records of the inventor’s interactions with AI. If your clients are relying heavily on AI to generate ideas rather than just refine text, they risk failing to meet the threshold for human inventorship.

The Public Disclosure Pitfall: Perhaps the most immediate risk you must warn clients about is the potential for inadvertent public disclosure. Prompting an AI tool with sensitive invention details can constitute a disclosure to a third party, potentially destroying patentability, waiving privilege, or losing trade secret protection.

Recent case law highlights this danger. In United States v. Heppner and Trinidad v. OpenAI, Inc., federal courts ruled that information sent to generative AI platforms constituted a public disclosure, resulting in the waiver of attorney-client privilege (Heppner) and the loss of trade secrets (Trinidad).

The terms of service were a crucial factor in both cases. The parties used consumer-tier accounts with terms-of-service agreements that allowed the platforms broad rights to use the inputted information.

Actionable Advice

To effectively guide your clients and protect their IP in the AI era, integrate the following advice into your client counseling:

• Encourage the inventor’s own words. Advise clients that short, manually drafted disclosures focusing on key details are more useful than lengthy AI outputs. Their own words allow you to recognize the true inventive features and think critically about the invention in ways AI can’t yet replicate.
• Audit their AI platforms. If a client insists on using AI, explicitly ask what tier of service they are using. Warn them against consumer-tier AI accounts. Advise them to strictly limit AI use to enterprise accounts or platforms with rigorous privacy protections and terms of service that don’t claim rights to user inputs.
• Repurpose AI as a refinement tool. Counsel clients to shift how they use AI. Instead of asking a chatbot to extrapolate a small idea into a large volume of text, advise them to use AI to draw out specific concepts or format their existing notes.
• Advise documentation of human contribution. In light of potential PTO rule changes, recommend that clients document their independent, human contributions to the invention before feeding those concepts into an AI tool for refinement. In many cases, this may simply be the prompt itself or documents provided along with a particular prompt.

AI tools promise significant efficiencies, but the risks at the invention disclosure stage are substantial. By proactively addressing these issues, IP attorneys can help their clients leverage technology without jeopardizing their patent rights.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law, Bloomberg Tax, and Bloomberg Government, or its owners.

Author Information

David Jordan is a principal at Fish & Richardson with experience in building patent portfolios across diverse technologies, including computer software, data analytics, and e‑commerce.

Benjamin Halkowski is an associate at Fish & Richardson who focuses on patent drafting and prosecution involving computer software, machine learning, and AI.

Write for Us: Author Guidelines