Chat GPT and Patent Law

[Warren Smith]

ChatGPT is shockingly BAD at patent law. I find this surprising.

 

Q:

What happens if party A files a patent application, and then party B files a substantially identical patent application prior to the publication of the patent application by party A?

 

A:

If party B files a substantially identical patent application prior to the publication of the patent application by party A, this can create a potential issue of double patenting. Double patenting refers to the situation where two or more patents claim the same invention or invention that is substantially the same.

 

If the US Patent and Trademark Office (USPTO) determines that the patent applications filed by party A and party B are substantially identical, the USPTO may reject the patent application filed by party A or require party A to take steps to overcome the double patenting rejection.

 

To overcome the double patenting rejection, party A may need to file a terminal disclaimer or a restriction requirement. A terminal disclaimer is a statement filed by the applicant that disclaims any patent term beyond the expiration date of a related patent or application. A restriction requirement is a requirement by the USPTO to elect a single invention for examination if the application contains claims to more than one distinct invention.

 

It is important to note that the determination of whether two patent applications are substantially identical is a factual and legal analysis that can be complex. It is advisable to consult with a patent attorney or agent if you are involved in a situation like this.

[Matthew Schlecht]

On Sat, Apr 8, 2023 at 7:07 PM Warren Smith <warren...@comcast.net> wrote:

 

It is important to note that the determination of whether two patent applications are substantially identical is a factual and legal analysis that can be complex. It is advisable to consult with a patent attorney or agent if you are involved in a situation like this.

Perhaps it's encouraging for ChatGPT to "realize" that, paraphrasing the words of (Dirty) Harry Callahan, "A bot's got to know its limitations." 

So, for now, patent lawyers need not worry about losing their jobs.

Then again, perhaps the creators have not yet included a substantial amount of legal content in ChatGPT's corpora diet. Properly fed, I would think that the bot would be quite useful going through case law, decisions, and judgments, and be able to present a brief at least as worthwhile as what a first-year law student could prepare.

Matthew Schlecht, PhD
Word Alchemy Translation, Inc.
Newark, DE, USA
wordalchemytranslation.com

[Joe Jones]

GPT is not the only game in town. There is a startup in Japan called IPDefine that is using AI to do prior art searches and find patent infringement, and charging patent holders and legal professionals a pretty penny to do so (but still less than it would cost to do these things manually).

---

From: hon...@googlegroups.com <hon...@googlegroups.com> on behalf of Matthew Schlecht <matthew.f...@gmail.com>
Sent: Sunday, April 9, 2023 11:54 PM
To: hon...@googlegroups.com <hon...@googlegroups.com>
Subject: Re: Chat GPT and Patent Law

 

--
You received this message because you are subscribed to the Google Groups "Honyaku E<>J translation list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to honyaku+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/honyaku/CALnYxczPpppJLPUNy5rYKtMG7E7W01HMOS7jqaRHMNK%2BJpFMCg%40mail.gmail.com.

[Matthew Schlecht]

On Mon, Apr 10, 2023 at 9:19 AM Joe Jones <j...@djjon.es> wrote:

GPT is not the only game in town. There is a startup in Japan called IPDefine that is using AI to do prior art searches and find patent infringement, and charging patent holders and legal professionals a pretty penny to do so (but still less than it would cost to do these things manually).

I'm not surprised, and we should expect more of this sort of thing.

It's just a short stroll from what copyright trolls do - they train bots to crawl the internet looking for images that are protected by license, and when they find one they issue a "pay-up" demand email.

I was caught up in one of these about eight years ago. I had included a picture of a DNA molecule in an ATA presentation, and then archived the talk in PDF form on my website. My faulty and superficial understanding at the time was that the image was in the public domain. Hah! Turned out that no licensing fee was required, but it was necessary to obtain permission before use. I negotiated a settlement and paid up.

The intriguing thing is that the bot discovered the digital signature of an image that had been embedded in a PPTX which was then converted to a PDF.

Those little bots can crawl into anything!