NEWS
AI Training, Fair Use, and Meta
We posted an article recently covering the Copyright Office’s report on AI training and fair use.
Just to review quickly, Fair Use is the provision of copyright law that allows for limited usage of copyrighted materials in certain conditions without needing to obtain prior permission from the copyright holder.
AI “training” refers to the specific data, including copyrighted works, from which the AI learns.
A recent lawsuit, filed by thirteen authors, accused Meta of copyright infringement. The judge in the case ruled that the plaintiffs had failed to show that Meta’s models produced outputs substantially similar to their copyrighted works. He noted that it “is generally illegal to copy protected works without permission” but found that they had not demonstrated a compelling argument that Meta’s use of the material caused market harm.
It should be noted that the court ruled in this instance that the usage of copyrighted material counted as fair use, but only because there was insufficient evidence of market harm. This was not a ruling on the broader issue of AI training. There is a separate claim that Meta may have unlawfully distributed the works, which was not dismissed and will proceed.
The judge rejected Meta’s argument that public interest justified unauthorized use of copyrighted material. Meta’s claim that barring usage of copyrighted material would halt AI development was called “nonsense”, and the judge added that if the books were as valuable to AI as Meta claims, they should find a way to license them.
While Meta won this particular case, the judge noted that stronger claims and better evidence could render a different outcome.
When Trying to Save Money May Cost More
The USPTO’s fee schedule offers discounts for small and micro entities.
Small entities are defined as a person, small business of no more than 500 employees, or a non-profit organization. Micro entities are defined as meeting the requirements of small entities, plus not being named on more than 4 previous applications and a gross income of less than $241,830, and is not under obligation to assign or license to an entity with an income exceeding the amount listed above.
If you qualify as a small entity, then there will be a 60% reduction on most patent related fees.
If you qualify as a micro entity, the reduction is 80%.
Given the fee reductions, it perhaps is no surprise that the USPTO has found applicants trying to fit themselves into those categories when they don’t qualify. The PTO recently implemented statutory penalties for false assertions or certifications of small and micro entity status. That fee will be “not less than three times the amount an entity failed to appropriately pay”. The Patent Office will send out notices of payment deficiency and an order to show cause as to why the fine shouldn’t be assessed.
Enablement and Prior Art: What Agilent v. Synthego Teaches Us About the Limits of Patent Protection
In the world of patent law, “enablement” is a critical concept. But what many don’t realize is that how much enabling detail is required depends heavily on what you’re trying to do—get a patent or invalidate one.
That distinction came into sharp focus in the recent case of Agilent Technologies v. Synthego, and it hinges on a deceptively simple but powerful principle: To secure allowance on a claim in a patent application, you have to enable sufficient embodiments to satisfy the scope of the claim. But to invalidate a patent claim, you only have to show prior art enablement of one embodiment that falls within the scope of the claim. The distinction isn’t new but it’s nice to see the rule clearly delineated.
The Two Faces of Enablement
Under Section 112 of the Patent Act, an inventor who applies for a patent must fully enable the entire scope of their claimed invention. This is to ensure that patent holders don’t gain broader exclusivity than their disclosure justifies. In other words, you can’t just describe one version of an invention and claim all possible variations—you have to teach the public how to make and use all of what you’re claiming.
That’s exactly what the Supreme Court reinforced in Amgen v. Sanofi. There, the Court invalidated Amgen’s patent because it covered a broad set of antibodies, but only disclosed a few actual working examples. The rest would require painstaking trial-and-error to figure out, which the Court found unacceptable.
But in Agilent v. Synthego, the Federal Circuit drew a sharp line.
When it comes to anticipatory prior art—that is, prior publications that can invalidate a patent—the rules are different. Prior art does not need to enable every version of the claimed invention. It only needs to enable one working embodiment.
A Research Proposal as Prior Art
Agilent’s patents related to chemically modified guide RNAs (gRNAs) that resist degradation in CRISPR gene-editing. Synthego challenged these patents by pointing to an abandoned patent application from Pioneer Hi-Bred. This old application described broad, theoretical possibilities for modifying gRNAs—but it:
- Contained no proven examples
- Proposed quadrillions of combinations
- Was ultimately withdrawn after failing to work
Agilent argued that this reference wasn’t enabling at all—it was more of a research proposal than a working invention. And they leaned heavily on Amgen, saying that just like in that case, the disclosure was far too vague and would require unreasonable experimentation.
But the court rejected this argument, and here’s why:
Amgen was about how much you need to disclose when you’re trying to get a patent.
Agilent was about how much is needed in prior art to block someone else’s patent.
The bar is simply much lower in the latter.
Why This Matters
By refusing to apply Amgen’s strict enablement standards to the Pioneer Hi-Bred prior art, the court effectively created a firewall.
This firewall draws a clear line between:
- Section 112 enablement: For your patent to be valid, you must enable the full scope of what you’re claiming.
- Prior art enablement: To defeat novelty, a reference must only enable a single embodiment—a single working version of what’s being claimed.
This distinction has profound implications.
The Power of Prophetic Prior Art
What’s especially striking is that the Pioneer Hi-Bred application was never proven to work and was ultimately abandoned. It consisted largely of prophetic examples, many of which turned out to be false or at least unproven.
But the court said that didn’t matter.
If even one of those prophetic examples was enabling—if a person of ordinary skill in the art could read the disclosure and reasonably make it work without undue experimentation—then the entire patent could be invalidated.
Key Takeaway
If you’re seeking a patent, courts will hold you to the high enablement standard of Amgen: you must teach others how to make and use your invention across its full scope.
But if you’re trying to invalidate someone else’s patent, even a speculative, abandoned patent application can do the job—so long as it enables one embodiment.
That’s the takeaway from Agilent v. Synthego:
In patent law, not all enablement is created equal.
Jazz Loses the Tune: Courts Strike a Sour Note on System Patents
In a decision that could ripple through the pharmaceutical industry, the Federal Circuit recently ruled on Jazz Pharmaceuticals, Inc. v. Avadel CNS Pharmaceuticals, LLC, clarifying the limits of injunctive relief and patent listings under the Hatch-Waxman Act. The May 6, 2025 ruling emphasized a key point: even when a company is found to infringe a valid pharmaceutical patent, courts can’t stop it from engaging in clinical trials or offering open-label drug extensions—so long as those activities fall under the Act’s “safe harbor” provision.
A Closer Look at the Case
The controversy centered around Jazz’s blockbuster narcolepsy drug, Xyrem, which contains GHB, a powerful sedative also known for its abuse potential. Jazz had patented a single-pharmacy distribution system for tracking and dispensing the drug. This system was at the heart of their patent dispute with Avadel, who sought approval for a competing GHB-based drug, FT218—but with a once-nightly dose and a different distribution method.
When Avadel submitted its New Drug Application (NDA), the FDA still required them to certify against Jazz’s listed patent. Jazz then sued for infringement. Avadel fired back, challenging the patent’s inclusion in the FDA’s Orange Book (a listing of patents related to approved drugs), arguing that Jazz’s patent didn’t actually claim a drug or a method of using one—which are the only categories eligible for listing.
The District Court Sides with Avadel
The district court agreed with Avadel, finding that Jazz’s patent claims were for a “system,” not a method of use or a drug itself. That distinction matters: only drug compositions and methods of use are valid for listing in the Orange Book. The court ordered Jazz to delist its patent, and Jazz appealed.
The Federal Circuit Weighs In
On appeal, Jazz argued two main points:
- That its patent did in fact claim a method of using the drug.
- That even if it didn’t, the FDA’s broader regulatory language should apply, allowing system claims to count as method claims.
The Federal Circuit wasn’t persuaded. It upheld the lower court’s reasoning, explaining that:
- Patent law requires method claims to involve specific steps. Simply describing a system—even one that enables a method—is not the same.
- The FDA regulation does not override that fundamental distinction in patent law.
- Jazz’s system patent couldn’t be reclassified as a method-of-use patent simply because it was related to conditions for prescribing the drug.
What About Clinical Trials?
An additional issue was whether Jazz could stop Avadel from continuing clinical research activities related to FT218. The appellate court clearly stated: No. Even when infringement is found, a district court cannot block clinical trials or open-label extensions if they’re protected by the Hatch-Waxman Act’s safe harbor—a provision designed to encourage innovation and expedite FDA approvals.
Why This Matters
This ruling reinforces a critical boundary between commercial infringement and regulatory research activities. It also tightens the standards for what can be listed in the Orange Book—potentially preventing brand-name drugmakers from using system patents to block competition unfairly.
For biotech and pharma companies, this is a strong reminder: if your patent doesn’t claim a drug or a method of using it, it doesn’t belong in the Orange Book. And for generics and 505(b)(2) applicants, it’s a win that helps preserve the streamlined pathway to market entry—even when innovative system-based solutions are at play.
PatChat is now operational
We have added some AI functionality to our website in the form of a personalized Chat GPT-based AI agent.
You can access the chatbot at the lower righthand corner of the website.
The AI is trained on Bob’s books, and the FAQs from our website, so asking it questions will be like talking to Bob himself.
Have fun with it, but don’t bother asking non-IP related questions, it’s trained to stick to IP related topics.
IP Updates in Korea
Korea has added exporting an invention to the list of what constitutes “practicing an invention”. This means patent holders can claim damages from overseas sales more actively.
Korea has also introduced two limitations on patent term extensions: a patent term can’t be extended over 14 years from the date of receipt of marketing approval, and only one patent becomes eligible for patent term extension per approval.
In trademark litigation, the Korean Supreme Court has clarified standards for consumer surveys, used as evidence of consumer recognition or perception.
USPTO discontinues the Accelerated Examination program
Given the long wait times for patent prosecution, the USPTO had offered several ways to speed the process up. One of those, Accelerated Examination, offered the option of having the attorney certify much of the examination work in advance. We used the accelerated examination process several times, and got quite good at it. In one instance we received notice of allowance in less than a month! Nevertheless, it seems many attorneys did a poor job of searching (intentionally or unintentionally), so the original premise of faster prosecution was faulty.
Track One, or Prioritized Examination, is still an option for speeding up the prosecution process.