About the Authors

  • The Authors and Contributors of "Patent Docs" are patent attorneys and agents, many of whom hold doctorates in a diverse array of disciplines.
2018 Juristant Badge - MBHB_165
Juristat #4 Overall Rank

E-mail Newsletter

  • Enter your e-mail address below to receive the "Patent Docs" e-mail newsletter.

Contact the Docs


  • "Patent Docs" does not contain any legal advice whatsoever. This weblog is for informational purposes only, and its publication does not create an attorney-client relationship. In addition, nothing on "Patent Docs" constitutes a solicitation for business. This weblog is intended primarily for other attorneys. Moreover, "Patent Docs" is the personal weblog of the Authors; it is not edited by the Authors' employers or clients and, as such, no part of this weblog may be so attributed. All posts on "Patent Docs" should be double-checked for their accuracy and current applicability.
Juristat #8 Overall Rank


« Petition for Writ of Certiorari filed in DABUS AI-as-Inventor Case | Main | Arbutus Biopharma Corp. v. ModernaTx, Inc. (Fed. Cir. 2023) »

April 09, 2023


Nice article Michael, and your point is well taken. The hype around ChatGPT2-4 reminds me of when the calculator, and later the home computer, came on the scene (yes, I was alive then).
These NLPs won't replace attorneys, just as neither the calculator nor the computer did. But I do expect them to make us more efficient. Just as computers freed us from having to look at the "pocked part," I believe NLPs like Chat GPT will free us from boolean searching case law.
Are you listening Lexis and Westlaw?

Very thought-provoking. My thoughts are that it is premature to test the skills needed to manage clients or a law firm in an examination of competence to practise as a lawyer. Here in Europe, the professions of i) patent attorney and ii) attorney at law, are different but of equal standing. To qualify as a patent attorney, since the 1970's (much longer in the UK) one must satisfy the examiners that one is competent to safeguard the interests of a client. One does this by answering questions in which one first studies the client's prototype, and the given prior art, and must then draft a set of claims that captures the full scope possible while at the same time satisfying all conditions of patentability, finding that narrow line that is not an inch too narrow nor an inch too wide. And then there is the exam paper in the UK, in which you get the client's patent, the prior art, and the accused infringing embodiment and are required as your answer to write an opinion covering i) infringement ii) validity, and iii) what the client can do to optimise their situation.

Examining the answers and deciding who passes and who fails is a tough task for the (human) Examiners. But how else shall you protect the public from patent attorneys who put out their brass plate but are not competent to deliver reliable advice?

One reads that AI will replace members of the learned professions before it replaces humans with the skills to care for the bodily needs of other humans. As a patent attorney, I'm doubtful about that.

I have to disagree with perhaps an unwitting premise of the story (as well as the asserted general acceptance that GPA is anything other than being a memorization 'tool' in its own right).

The 'evolution' story here is NOT mere advance of a tool being used, but the 'tool' itself being unleashed on tasks of creativity to which no human 'in the loop' satisfies a legal definition of being the 'inventor' of the output of that loop.

Yes, I "get" that a spit-back memorization is a bit of a non-sequitur to the larger point of the intersection of Intellectual Property Law and the continued development of AI (at pure memorization - I have to scoff at ONLY achieving at the 90th percentile...?).

It is a disservice to dance in the weeds. We've already had several years since the DABUS case first came on the scene, and putting off serious discussion of the impacts of non-human innovation serves no one.


I am similarly not a fan of GPA, but it is slightly more helpful than just regurgitation-based standardized examination. Evaluating humans is hard but many of the ways we do it now are useless at best and harmful at worst.


I agree Dr. Borella.

Evaluating more than rote, practically necessitates nuance and context, and risks losing objectivity (interactions with the person doing the evaluation may well become more than passively interactive).

Add in the desire for repeatability and low cost and quick turnaround while desiring to maintain accuracy....

Back in the dark ages, 1978, when I took the patent bar, it was a fully written exam, requiring drafting of claims/responses/client letters. Some memorization helped, for sure - the more you had in mind, the faster you could pick out the issues you needed to deal with and address them. Now and for some time it has been multiple guess - easier for the examiners, but far less helpful in testing competence.
And if you pass, you can hang out your shingle, even if you've never seen a patent application.
At least some other countries require an "apprenticeship" before you get licensed (cf. Max Drei's comment above).

> Aspiring lawyers typically spend 10-12 weeks taking a bar exam review course, which involves committing massive amounts of legal rules and principles to memory, as well as learning how to write essays in a formulaic fashion (IRAC).

If one wants to ascribe this to malice, rather than to incompetence in assessment design, one might note that the "10-12 weeks of study" and utility of expensive "bar exam review courses" add a distinctly economic aspect to who is most able to pass the bar, while requiring nothing more than the ability to memorize and operate in a formulaic manner.

It's not unreasonable to suspect that the bar exam, as it exists now, is designed primarily to prevent less-privileged people from passing (and thus gaining access to a middle-class income) while still allowing the various flavors of Bush, Kennedy, and other famous idiot children of wealth to get a credential by tossing enough money and time at the problem.

At least some of the famous children had to take the exam multiple times. John Kennedy Jr. and Richard M. Daley needed three tries each.

Ascription to malice is unbecoming.

Many of us (myself included) belong to that purported "victim" class of "less-economically-well-off."

I passed the bar with my first attempt without the expensive bar exam review courses.

Of course, the bar IS set up to exclude and/or limit membership in 'the club' - the guild model (harkening back to feudal times) is ALL ABOUT controlling supply and demand. Sure, some 'token' arguments can (and will be appropriately) made as to controlling quality, but only the most niave do not consider the larger guild controls as providing very important reality 'checks.'

The comments to this entry are closed.

May 2024

Sun Mon Tue Wed Thu Fri Sat
      1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30 31