Intellectual property attorney Damien Riel and computer programmer Noah Rubin had an idea while talking in a hotel bar. What if computers could be used to create all sorts of musical melodies in the same way that hackers use “brute force”? How to run all possible combinations of strong passwords?
That night they built a prototype program to create 3,000 melodies. And so the “All the Music” project began. Melody soon he increased to 68 billion and as of today he has reached 417 billion. The melody is copyrighted and made available to the public, ensuring that songwriters cannot inadvertently infringe existing music rights and avoid frivolous or opportunistic lawsuits.
Riehl and his collaborators then realized that the model could also be applied to patents, and began work on a follow-on project called All the Patents.
“What this project will do is take all the claims of all the patents ever filed, use a machine to recombine them, and then look at all the possible claims of all the previous patents. It’s about creating every possible combination,” says Leal. Once published, they act as “prior art.” This is the patent law concept that an invention cannot be patented if it has already been described, is obvious, or is not new.
Its purpose is to modify existing patents for purely strategic purposes, i.e., with no intent to develop the invention, but to deter other companies from developing similar innovations or to gain some financial benefit. to prevent companies from generating thousands of new patents on their machines.
Lille’s project means ‘denying the possibility of recombination of old things’. [for a patent] It’s so expensive that it’s not worth the expense,” he explains. “Instead, we need to innovate with products, not patents.”
Both efforts show how artificial intelligence can be applied to intellectual property to support or impede innovation.
Although AI cannot currently be named as an inventor on most patents, it can speed up both the innovation itself and the drafting and filing process. “When it comes to drafting, as in many other applications, AI can help create text,” said Stanford University Law School Professor Mark Remley. “And it speeds up the process and makes it cheaper.” But he says, “There are tools out there that let you create the patent claims you want to own, and AI will draft the actual text of the envisioned technology.” ” he added. It is contrary to the original mechanism of patent law. “
The use of generative AI has increased significantly since the launch of the ChatGPT chatbot in November 2022, but older forms of AI have been used in the patent space for over a decade. One of his first examples was Lex Makina, a software that uses machine learning to examine data sources and provide legal analysis. It was spun out of Stanford University and was founded in 2006 by Lemley, George Gregory, and Joshua Walker.
Walker, an IP attorney and AI consultant, said that engineers and computer scientists are interested in how their inventions are patented, and not only are they more likely to create such solutions, He said that due to the openness of the information, there has been considerable progress in obtaining patents for the use of AI. And patents are associated with enormous economic value.
At Toyota North America, one of the busiest patent applicants in the US, the team under IP management consultant Gunnar Heinisch was using AI for years before ChatGPT came along. “We have used algorithms and machine learning to assess patent quality,” he says. “We can assess potential ideas, what are good inventions to consider, and whether they are patentable.” He hopes to be able to summarize patent applications in plain English for readers of
In the long term, AI may replace much of the human labor involved in the patent process. This requires rethinking not only how patents are drafted and valued, but also the role of patent attorneys and other professionals.
“Patenting is getting more complicated,” said Vaseharan Kanesaraja, head of intellectual property strategy development at analytics firm Clarivate. “And the traditional model of patent attorneys being single-disciplinary experts is crumbling.”
