By Courtney L. Baird and Ryan S. Crawford
In an issue of first impression, a federal court held that information a defendant input to a consumer generative AI system on his own initiative is not protected by the attorney-client privilege or the work product doctrine. That holding extended to documents the defendant generated using AI and later shared with counsel.
By Sharon Caffrey and Seth H. Dawicki
Duane Morris Partner Sharon L. Caffrey and Associate Seth H. Dawicki co-authored the article “AI Transcription Tools: Privacy, Privilege and Ethical Pitfalls” for Law.com. Sharon and Seth discuss the privacy and ethical pitfalls of the widespread adoption of artificial intelligence transcription tools.
Read a reprinted version of the article on the Duane Morris LLP website.
Artificial intelligence is reshaping how patent attorneys approach prosecution work. From prior art searches to claim drafting, AI tools promise efficiency gains that were unimaginable a decade ago. But with these promises come legitimate questions about reliability, risk, and the boundaries of responsible use.
This five-part blog series examines AI’s role in patent prosecution through multiple lenses. Rather than offering a simple thumbs-up or thumbs-down verdict, we explore the mainstream consensus, the contrarian arguments, and the evidence that would settle the debate. Whether you’re an AI enthusiast or a skeptic, this AI-Assisted series will give you a framework for thinking critically about these tools.
Efficiency gains, blind spots, and what evidence would actually settle the debate.
The Promise
Every patent practitioner knows the frustrations associated with prior art searching. You’re hunting through thousands of patents, non-patent literature, foreign filings, and technical publications, all while the clock runs and the client asks when they can file. AI tools promise to transform this bottleneck into a streamlined process, surfacing relevant references in minutes rather than days.
The mainstream view holds that AI excels at the initial sweep. These tools can ingest an invention disclosure, identify key technical concepts, and return a prioritized list of potentially relevant documents. For freedom-to-operate analyses, AI can map patent landscapes, flag blocking patents, and help practitioners understand where the white space lies. Tasks that once consumed twenty hours of associate time might now take two hours of AI-assisted partner time including review.
The efficiency case is compelling. One attorney can handle the volume that previously required a team. Clients get faster turnarounds. And firms can offer competitive fixed-fee arrangements, with higher billing rates but without sacrificing margins.
The Pushback
Not everyone is convinced. Critics raise three substantive objections.
First, there’s the insufficiency argument. AI tools operate on pattern matching and semantic similarity. But finding invalidating prior art often requires the kind of lateral thinking that comes from deep technical expertise. A mechanical engineer might recognize that a hydraulics patent from an unrelated industry anticipates a client’s automotive innovation, a connection that depends on understanding the underlying physics, not just keyword overlap. Can AI make those leaps?
Second, practitioners worry about systematic blind spots. AI searches are only as good as the databases they access and the algorithms that power them. What about foreign-language references that haven’t been translated? Non-patent literature from obscure trade publications? Unpublished theses? A skilled human searcher knows to look in unexpected places. AI may not.
Third, and perhaps most troubling, is the false confidence problem. When a partner receives a neatly formatted AI search report, there may be a psychological tendency to trust its completeness. The very professionalism of the output may reduce the scrutiny it receives. Paradoxically, AI-assisted searches might get less rigorous human review than purely manual searches did in the pre-AI era.
What Would Settle the Debate?
Proponents of AI-assisted searches need to produce controlled studies comparing outcomes. If AI-assisted prior art searches identify invalidating references at rates equal to or better than traditional searches, with documented time savings, the efficiency case becomes encouraging. Data from post-grant proceedings would be particularly persuasive: do patents that underwent AI-assisted prosecution survive IPR/PGR/REEXAM challenges at comparable rates?
Skeptics, meanwhile, would point to malpractice claims, post-grant invalidations, or prosecution failures traceable to AI search gaps. If patterns emerge showing that AI systematically misses certain categories of prior art, that’s evidence the technology isn’t ready for unsupervised deployment.
Until that evidence accumulates, prudent practitioners will likely treat AI as a powerful but humble servant that is a supplement to, rather than replacement for, human expertise in prior art searching.
California Gov. Gavin Newsom has historically vetoed legislation surrounding automation. However, as he sets his sights on the White House, Newsom has been slow to set new regulations surrounding artificial intelligence while facing mounting pressure from the state’s unions. Duane Morris Partner Alex Karasik discussed the legal risks California employers should keep in mind when implementing the new technology.
Read the full article in International Employment Lawyer.
The USPTO’s new director has singled out AI, distributed ledger technologies and diagnostics as prime areas of innovation that merit patent protection. Companies, investors and other stakeholders are closely watching how the USPTO’s active guidance may better align patent practice with the ingenuity and societal benefits these technologies represent.
Last week, Duane Morris kicked off a new multipart webinar series—What’s Watt— taking a deep dive into the critical relationship between energy and modern data centers and highlighting the trends and technologies reshaping digital infrastructure. The series launched with a state-of-the-market discussion with DigitalBridge’s Jeff Ginsberg and Duane Morris’ Robert Montejo, moderated by Brad Molotsky. The panel offered insights on the industry’s key trends.
Here’s Watt you missed:
1. Power Is Key
Watt’s Old: Smaller-scale developments with traditional grid access.
Watt’s New: A mix of utility power, onsite generation, and creative energy strategies to meet a much greater demand.
AI’s explosive growth is reshaping the energy landscape, driving an unprecedented need for reliable, large-scale power. Training and operating advanced AI models requires massive compute clusters that draw far more electricity than traditional cloud workloads, pushing data centers into power ranges once associated with heavy industry. As organizations race to deploy AI capabilities, the demand for high-density facilities, fast interconnection, and resilient energy infrastructure is outpacing what many utilities can deliver on typical timelines. This surge is forcing developers, operators, and policymakers to rethink how and where digital infrastructure is built—prioritizing power availability, alternative generation sources, and innovative grid partnerships to keep pace with AI’s accelerating requirement.
2. Infrastructure Is Expensive
Watt’s Old: Traditional loan structures with shorter maturities.
Watt’s New: Large-scale, multilayered financing with longer terms and institutional investors capable of absorbing significant risk.
The next wave of data center development—driven by AI-scale power and capacity requirements—will require hundreds of billions of dollars in capital. Whether this growth ultimately forms a bubble remains unclear, but its scale is already reshaping credit markets and stretching the capacity of conventional lenders. As banks reach concentration limits and face regulatory constraints, developers increasingly rely on institutional investors, sovereign funds, infrastructure platforms, and hyperscalers with trillion-dollar balance sheets to support long-duration projects. These deals frequently involve complex capital stacks, special purpose vehicles, and financing horizons of 15–25 years to match the lifecycle of large campuses and energy assets.
3. Focus on Execution
Watt’s Old: Out of sight, out of mind.
Watt’s New: Plan deliberately and anticipate environmental, regulatory, and community challenges.
Ambitious AI-driven demand has raised the stakes for planning and execution in large-scale data center and energy projects. Power generation—whether grid-supplied or onsite—introduces thermal loads, water requirements, land-use impacts, and transmission needs that must be addressed early to keep projects viable. In many regions, particularly in the Western U.S., water constraints, aquifer depletion concerns, and limited cooling alternatives can quickly challenge site feasibility. Local infrastructure pressures, such as noise, construction logistics, easements, and grid constraints, often converge with environmental and community concerns, creating conditions ripe for pushback or organized resistance. Effective execution now means proactive engagement, rigorous resource planning, and transparent mitigation strategies to avoid delays and ensure that projects scale responsibly amid real and growing demand for AI infrastructure.
Find a recording of the full-length discussion here.
Watt’s Next: Our webinar series continues with What’s Watt: Nuclear Power and the Future of Data Center Construction
This year has been a busy one in the generative artificial intelligence (gen AI) class action litigation landscape. New pleadings were filed, including several new class actions, several consolidated and amended complaints, and one appeal. Several key decisions were issued, including a trio that formed a three-way split of authority on how to determine whether training a gen AI model on copyrighted materials constitutes “fair use” under the Copyright Act. Additionally, one humongous settlement was reached.
Read the full article by Justin Donoho.
By Mark Lerner
Following a refusal to grant a copyright registration to Stephen Thaler for a work whose sole author was identified as “Creativity Machine,” a generative AI Thaler created, the D.C. Circuit affirmed that works authored exclusively by artificial intelligence are ineligible for copyright protection under the Copyright Act, which the court read to require human authorship, in keeping with the Copyright Office interpretation and prior case law. A petition for certiorari and a supporting amicus brief now ask the U.S. Supreme Court to take up the question of whether the Copyright Act requires human authorship, arguing that the statute’s text, structure and purpose do not categorically impose such a requirement and that existing doctrines leave room for AI to be recognized as the author of protected works.
Read the full Alert on the Duane Morris website.
Duane Morris will host the third session of its Wearable Webinars Series, Product Liability and IP Strategies for Wearables, on Tuesday, November 4, 2025, 12:00 p.m. to 12:30 p.m. Eastern.
Agatha Liu, Ph.D., will cover how wearables with diagnostic, monitoring or therapeutic claims fall under the FDA’s software as a medical device framework, including predetermined change control plans, good machine learning practices and real-world performance monitoring of adaptive algorithms.
By Alex. W. Karasik, Brian L. Johnsrud, and George J. Schaller
Duane Morris Takeaways: On October 13, 2025, California Governor Gavin Newsom, issued a written statement declining to sign Senate Bill 7 – called the “No Robo Bosses” Act (the “Act”). While the Act aimed to restrict when and how employers could use automated decision-making systems and artificial intelligence, Governor Newsom rejected the proposed legislation in terms of the Act’s broad drafting and unfocused notification requirements. Governor Newsom’s statement reflects an initial rebuttal to a wave of pending AI regulations as states wrestle with suitable AI guidance. Given the pro-employee tendencies of Governor Newsom and California regulators generally, this outcome is a mild surprise. Employers nonetheless should expect continued scrutiny of AI regulations before enactment.
This legislative activity surely sets the stage for what many believe is the next wave of class action litigation.
See more on the Duane Morris Class Action Defense Blog.
