Don't Believe the Hype
The USPTO is going to put some new applications through an AI search and give applicants “an initial communication identifying a ‘top ten list’ of potential prior art issues in need of attention.”
- Too soon. I don’t have confidence in an unsupervised AI tool like this. The technology isn’t there yet and will likely give applicants, especially pro se applicants, some misleading and confusing results along with the helpful ones.
- Wrong stage. The best use of this kind of USPTO AI tool would be to have it available for applicants to use BEFORE filing, since once something is filed, the scope of allowable changes is limited.
- Doesn’t replace a patentability search. This doesn’t replace having a thorough patentability search done before moving forward with an application, and I’m concerned laypersons will think otherwise.
- Patent examiners think it’s a problem. From what I’m hearing from patent examiners, it sounds like (a) they think this AI tool isn’t very good yet and (b) that it is the first step in transitioning eventually to largely AI examination. Once again, the AI technology is absolutely not there yet. Can it help? Sure. Can it replace a human examiner? Not even close. (N.b. my master’s degree concentrated in AI, natural language processing, and LLMs.)
- Pay to play. Of course, there is a fee to pay to participate in the pilot program.
- Propaganda. Don’t buy that line about the USPTO Director’s “commitment to dramatically improve examination quality.” What he’s doing to examiners behind the scenes is absolutely counterproductive to that goal.
This program could be useful to some applicants, but it’s definitely overpromising and trying to draw attention away from the chaos and poor treatment of examiners happening at the USPTO. It’s almost like the leadership is trying to break it so it can be turned over to a private company to handle. I’m NOT happy about my clients paying the same fees for poorer quality examination, and that’s exactly what’s going to happen thanks to new policies about examiner quotas and limits on their ability to help applicants.
The Text of The USPTO Press Release:
ALEXANDRIA, VA—In furtherance of the U.S. Patent and Trademark Office’s (USPTO) steadfast commitment to dramatically improve examination quality, response time, and efficiency, the agency today announced the launch of its Artificial Intelligence Search Automated Pilot (ASAP!) Program. The ASAP! Program will test the USPTO’s internal artificial intelligence (AI) tool’s efficacy for conducting pre-examination prior art searches.
“To the Senate Judiciary Committee and inventors of all stripes, I promised to ‘lean-in’ to AI,” said John A. Squires, Under Secretary of Commerce for Intellectual Property and Director of the USPTO. “Our ASAP! announcement today is the first of many planned AI pilots designed to help examiners and applicants alike ensure their patents are born strong. Quality starts at filing—and our enterprising examiners have developed a cadre of new tools that re-imagine workflows and attack thorny chokepoints that constrain productivity, throughput, and ultimately quality for new applicants. ASAP! is intended to foster common ground in the first instance. This is us leaning-in. Watch this space.”
The new pilot will provide applicants with an initial communication identifying a “top ten list” of potential prior art issues in need of attention. By generating an AI-Assisted Search Results Notice (ASRN), ASAP! will afford applicants an early opportunity to assess their claims against the prior art before substantive examination. Indeed, the pilot will look to inform applicants of potential courses of first action responses such as:
- Filing a preliminary amendment;
- Marshaling evidence for affidavit practice and notice taken by the USPTO;
- Requesting deferral; or
- Filing petitions, including express abandonment, to seek a refund of certain fees if examination is no longer desired.
The USPTO will conduct its automated search using an internal AI system that derives contextual information from the classification of the application under the Cooperative Patent Classification (CPC) system, as well as from the specification, claims, and abstract of the application.
The USPTO will utilize the pilot results to assess outcomes for pre-examination searches, evaluate the scalability of generating the ASRN, and collect data to inform next steps and companion pilots. To participate in ASAP!, applicants will be required to file a specific petition accompanied by a petition fee.
https://content.govdelivery.com/accounts/USPTO/bulletins/3f63793
RSS - Posts
