USPTO conference on Artificial Intelligence and IP: a report

Somewhat belatedly, the IPKat is delighted to publish the following post from Katfriend Lee Curtis on the conference held at the United States Patent and Trademark Office in Alexandria, Virginia on January 31, 2019, entitled “Artificial Intelligence: Intellectual Property Policy Considerations”.

In contrast to many other conferences on AI, this gathering concentrated on the impact of Artificial Intelligence on how intellectual property law will operate in the years ahead, as opposed to how AI will impact on attorney firm practice. There were 28 speakers, with six panels spanning all forms of IP, including patents, designs, copyright and trade marks. The full agenda of the conference can be found here. The main issues addressed at the conference can be summarized as follows:

• Economic frameworks and impacts: How will AI impact how we innovate and create, produce and deliver new products and services, as well as how we work and engage with one another? How might the new tools and insights offered by AI improve our ability to recommend and assess changes to IP policy?

• Patents and Trade Secrets: How can AI-related inventions be protected? What are the challenges that inventors and rights-holders face, whether they are a Fortune 500 company, a start-up, or a non-profit?

• Trademarks: Will AI change the likelihood of confusion and liability? How will it impact the branding of products and the protection of trademarks?

• Copyright: Who is the author of AI-generated content? Are such works copyrightable? What policy implications arise from the use of copyrighted works for the purposes of machine learning?

• IP enforcement: Counterfeit goods make up an estimated $461 billion or 2.5 percent of all global trade. How is AI improving counterfeit detection? How can we leverage new technologies to solve this age-old problem?

• International perspectives: How are other major economies addressing AI, and in what ways do they differ from the approach taken by the U.S? To what extent do these differences matter to U.S. companies and researchers?

The opening remarks were provided by Andrei Iancu, Under Secretary of Commerce for Intellectual Property and Director of the USPTO and Lynne Parker, Assistant Director for Artificial Intelligence, White House Office of Science and Technology Policy. Director Iancu provided a summary on the importance of AI for the US and global economy and the importance of its impact on IP law.

He further summarized some of the initiatives being implemented by the USPTO regarding AI. They include a new cognitive assistant called “U” or “Unity,” which is a tool that is intended to allow patent examiners, through a single-click, to conduct a “federated search” across patents, publications, non-patent literature, and images and through AI and machine learning-based algorithms, which would then present to the examiner the results in the form of a “pre-search” report.

The USPTO is also exploring semi-automated tools for “search query expansion,” trained to mine technology-specific synonyms with the help of crowd or “examiner-sourcing.” As well, the USPTO is testing new AI tools and techniques, such as robotic process automation (RPA) that could generate smart office action templates, which are automatically populated based on the interactions between examiner and attorney. This will save the examiners from some of the more tedious clerical aspects of generating office actions.

Ms Parker provided a useful and interesting summary of what is meant by the term ‘artificial intelligence.’ In so doing, she framed the subsequent discussions and provided an important staring point in any discussion of AI.

Focusing on the trademark panel, the participants were Dana Brown Northcott, Associate General Counsel, Intellectual Property, Amazon.com, Inc., Makalika Naholowaa, Senior Attorney and Head of Trademarks, Microsoft Corp, Adam West, Head of Marketing, Satalia in London and Susan Anthony, Attorney-Advisor, Office of Policy and International Affairs, USPTO, and Lee Curtis.

The trade mark panel discussion ranged over a broad range of subjects. Susan began by raising the question of whether AI has a branding problem. Adam West commented that we naturally overestimate the impact of any new technology, such as AI, in the short term, but underestimate its long term impact. How will AI impact on the way products are purchased, which after all is the basis of trademark law.

Makalika Naholowaa structured her answer in two layers, the public’s perception of AI in general and then the public’s perception of the impact of AI on marketing specifically. She emphasized the importance of clarity in educating the public and highlighting to them the benefits of AI. She commented that the ‘fourth industrial revolution’ is coming quickly on the heels of the ‘third industrial revolution’ linked to the internet. Most of us were around for the third industrial revolution, in contrast to the first industrial revolution from several centuries ago.

Makalika Naholowaa also observed that the way information in the purchasing decision is presented to the consumer may change with the introduction of AI. However, the information available to the consumer should not be reduced. Further, Makalika highlighted that AI will make existing processes faster and cheaper, providing direct benefits to consumers.

The discussion also considered whether some of the current terminology used in trademark law-- such as average consumer, likelihood of confusion, imperfect recollection and the slurring of trade marks-- may become redundant. After all, trademark law is based on human frailty. Can AI be confused? What happens when you take the ‘human’ and the ‘frailty’ out of the purchasing process, how does that impact trademark law?

Further, who is liable when an AI assistant purchases infringing product? Some guidance may be found guidance in EU decisions such as Google France and the eBay vs. L’Oréal decisions , which suggest that AI providers will not be found liable without ‘active’ knowledge of infringing activity.

Also discussed was how AI can help in the fight against infringers and counterfeiters with Amazon increasing efforts to make sure that counterfeiters do not even get to the detail pages. Amazon is developing AI machine learning at the backend to identify fraud vectors, ‘scooping’ behind the scenes unseen by the consumer to identify products which could be counterfeit and remove them from the detail pages. Amazon has already set up the Brand Registry as part of those efforts.

Dana commented that the fight against counterfeits will be a partnership between AI, academics, consumers and technologists and she also highlighted the move to product searching by voice. It is predicted by some commentators that within five years, 50%t of product searching will be conducted by voice, as a result of the increased use of AI assistants. This means that the phonetic comparison of trade marks will have an increased importance in trade mark comparison. This may also mean that some of burden may be pushed back onto the consumer to concentrate on the differences between brands.

A video recording of entire conference can be found here (around seven hours split into parts).

Picture on lower left of the Oracle at Delphi is in the public domain.