So, is the U.S. Patent and Trademark Office (USPTO) actually going to revolutionize IP practice with its new AI tools, or is this just another round of Silicon Valley buzzwords dressed up in government beige? Having spent two decades watching tech trends crash and burn, I tend to lean toward the latter, especially when the whiff of corporate-speak hangs heavy in the air. The USPTO’s recent pronouncements on AI-assisted inventions and their new search pilots are certainly getting a lot of airtime, but let’s cut through the static and see what’s really going on.
Inventorship: A Moving Target, or Just Dizzying?
The USPTO’s stance on who—or what—can be an inventor in the age of AI has been, to put it mildly, a rollercoaster. First, there was the Thaler ruling, which firmly established that AI can’t be an inventor. Simple enough, right? Then, in early 2024, the USPTO dropped some guidance that seemed to flirt with treating AI like a co-inventor, using a test designed for human joint inventors to evaluate an applicant’s interaction with an AI tool. The reaction? Predictably, a lot of head-scratching and alarm bells from practitioners who understood what this could mean down the line in litigation. Flash forward to November 2025, and Director Squires issues revised guidance, rescinding the previous one. Poof. AI is now just a tool, like a fancy microscope or a well-stocked digital library. Conception—the traditional legal standard—is king again. The USPTO won’t poke around unless you literally list an AI as an inventor. Easy peasy.
But here’s the kicker, the part that makes my cynicism antenna twitch: This revised guidance, while binding for examiners at the USPTO, doesn’t bind the courts. District courts, the Patent Trial and Appeal Board (PTAB), the Federal Circuit—they all still have the final say. The fundamental question of what “conception” really means when AI is doing the heavy lifting is still very much up in the air, and that’s where the real battles will be fought. The USPTO’s internal rules might change daily, but the courts’ interpretations are what’ll eventually matter, and those are far more stable—and expensive to figure out.
ASAP! and the Search for Actual Value
Beyond inventorship, the USPTO is pushing AI into the examination process itself. The Artificial Intelligence Search Automated Pilot Program, or ASAP!, is the shiny new object here. The idea is simple: applicants get AI-generated prior art search results before the examiner even looks at their application. Sounds efficient, right? It uses AI to sift through patents and publications, spitting out up to ten ranked documents. The goal? To help applicants make smarter decisions early on—amendments, dropping the application, whatever—before the first official office action.
Now, the USPTO has been fiddling with AI search tools internally since at least 2022, helping examiners find related patents beyond simple keywords. But ASAP! is the public-facing, applicant-facing version. And here’s where things get interesting. The numbers aren’t exactly setting the world on fire. As of mid-April 2026, only a fraction of the available slots in the program have been filled. The USPTO even waived the $450 petition fee and extended the deadline to try and drum up more interest. When your own fee waiver and deadline extension are necessary to get people to sign up, that’s usually not a sign of overwhelming success.
Look at what practitioners are saying. One attorney reported that the search results were barely relevant. Another noted that examiners have been telling management the AI search is ineffective, and management seems… indifferent. A third pointed out a critical lack of transparency: not enough has been disclosed about the methodology behind these AI searches. This isn’t just about a few bad results; it’s about a fundamental question of trust and utility. Are these tools actually making the process better, or are they just adding another layer of complexity and potential for error that will eventually lead to more disputes?
The Real Money Play
So, who actually benefits here? The USPTO gets to look like it’s “modernizing,” which plays well with Congress and the public. For practitioners, the promise is efficiency. But if the AI search results are junk, and the inventorship guidance is going to keep shifting based on court decisions, then what’s the real value? It’s easy to get swept up in the AI hype, but the folks who will ultimately profit are the ones selling the AI systems and the consultants who help navigate these ever-changing rules. The rest of us are left sifting through the fallout, hoping the government’s tech experiments don’t create more problems than they solve.
🧬 Related Insights
- Read more: Suzie Law: Open Source AI for Lawyers [Legal AI Evolution]
- Read more: AI Bias and Discrimination Law: Legal Risks of Algorithmic Decision-Making
Frequently Asked Questions
What does the USPTO’s ASAP! program do? ASAP! provides applicants with AI-generated prior art search results before their patent application is formally examined, aiming to inform early strategic decisions.
Will the USPTO’s AI inventorship guidance change again? The USPTO has already rescinded and revised its guidance once. Future court decisions could also impact how inventorship is interpreted, suggesting continued flux.
Is AI being used for patent searches at the USPTO? Yes, the USPTO has internal AI search tools for examiners and a pilot program, ASAP!, for applicants to receive AI-generated search results.
