Login | March 29, 2024

AI at IP at UA Law

San Diego-based AI startup founder Navid Alipour. Alipour heads the venture formation fund Analytics Ventures LLC. (Photo by Richard Weiner/Legal News)

RICHARD WEINER
Technology for Lawyers

Published: March 15, 2019

The 21st annual University of Akron School of Law’s Symposium on Intellectual Property Law and Policy featured the director of the United States Patent and Trademark Office (USPTO) and the scientist who led IBM’s Watson patent team, among other very distinguished panelists.

In his keynote address, Andrei Iancu, undersecretary of Commerce for Intellectual Property and director of the USPTO spent a considerable amount of time praising the university and the city for their historic contributions to the invention and development of new products.

He then touched on his work in trying to streamline the work of the USPTO and his attempts to create a more favorable system for creating legal protections for the ownership of new products. But in answers to several questions, he refused to speculate on how some cutting-edge products, including computer programs and artificial intelligence, would be affected by the patent process going forward.

At issue was a seeming clash between some recent Supreme Court and other federal court decisions that seemed to restrict inherent patentability and the need for inventors to protect ownership of their inventions advocated by Iancu and numerous patent attorneys in attendance.

This very esoteric discussion is well outside of the purview of a tech column, but one panel in particular on artificial intelligence did fit inside the legal tech field.

That clash is particularly relevant to the most advanced forms of computing, and especially the various kinds and applications of artificial intelligence.

The panel consisted of several patent attorneys, as well as IBM’s Scott N. Gerard, who led the Watson patent team, and San Diego-based AI startup founder Navid Alipour. Alipour heads the venture formation fund Analytics Ventures LLC. (Poetic justice moment—Alipour’s flights were delayed several times due to winter weather).

Gerard spent his allotted time discussing the various types of AI, deep learning (DL) and machine learning (ML), while patent attorney Nick Transier made the point that recent court decisions have made it difficult to inject predictability into the AI patent application process—to the point that he felt that there had to be a legislative solution.

The panel explained that AI is basically “data in—cool stuff out.” What happens in-between may or may not be patentable, depending on how original or common that the interior algorithms are. But they seemed to say that getting the necessary detailed formulae explained in a way someone outside the field could understand it is a difficult task. The formula for that is: (geeks--equal sign with a slash through it--English).

All of that of course is relatively moot unless the product can be monetized in the first place, which is where Alipour came in.

AI in the legal field can be divided into two distinct areas: Legal applications of AI (which seem to be primarily in e-discovery for now, although there are legal advice bots), and the laws surrounding monetizing AI itself like patent and copyright laws.

According to the International Data Corporation, AI products are projected to earn amongst $60 billion by 2025. Between 2015 and 2018, over 9,000 AI-based patents were applied for in the US alone.

This column has always been heavy on user interface and light on underlying laws, but I do talk about business models, and no product comes to market without funding.

Alipur pointed out to me that there is very little funding for a product that can’t be patented outside of academia, where most products are developed open source and made available to anyone.

Academia is where most computer and AI research begins, and many academics literally have no interest in monetizing their products, he said. That makes the whole idea of patenting and monetizing AI a really fluid, ever-changing process.

Alipour’s fund currently concentrates on medical AI. The most successful of the fund’s projects is CureMetrix, which successfully predicts breast cancer up to six years before any other testing has, Alipour said.

In talking about patenting the technology of the startups that his fund is backing, Alipour said that CureMetrix has become successful enough to be able to afford its own in-house counsel, but that the rest of the necessary legal work, including patent work, is done by outside counsel.

Regardless of the bumps and roadblocks, Alipour is highly optimistic that AI is a field that will be increasingly monetizable. It is the future, and the people at this conference were right at the edge of it all.


[Back]