Jared Council reports in the Wall Street Journal:
The U.S. patent office stated that applications wouldn’t be considered unless the applicants listed the inventors by their legal names. A response requested that the office recognize the (AI) inventor because there was no human inventor. The patent office is working to formulate policy positions on the issue. The judicial and legislative branches will have to decide whether patents can recognize AI systems as inventors. “I’m not sure we have AI yet where machines are doing the inventing. But assuming we will, I we want to make sure our laws provide clarity for inventors, owners, investors, so there are not clouds of uncertainty to slow down innovation.”
If an artificially intelligent system creates a new product, should patent offices recognize it as the inventor?
That’s the question at the center of a case making its way through patent offices in the U.S., Europe and the Middle East, one that business leaders and lawyers say is posing fundamental questions that could alter how centuries-old patent systems around the world operate.
The case involves an AI expert, a professor and a group of attorneys who filed two patent applications over the past year, designating an AI system as the inventor behind each. But the patent laws in some of the jurisdictions where the applications were filed only recognize “natural persons” or “individuals” as inventors.
The inventions are a specially shaped container lid designed for robotic gripping and a flashlight system for attracting human attention in emergencies. The group says both were created by an AI system called Dabus (short for “device for the autonomous bootstrapping of unified sentience”), which in turn was built by Stephen Thaler, founder and chief executive of Imagination Engines Inc. of St. Charles, Mo.
Mr. Thaler has spent at least a decade developing Dabus, which was built to ingest data about a range of subjects—including fractal geometry and flashing light patterns—and conceive ideas for products it hadn’t seen before. The group says Mr. Thaler has no background in developing container lids or flashlight systems, didn’t conceive of those two products and didn’t direct the machine to invent them—so it’s improper to list him as the inventor.
Ryan Abbott, a law and health-sciences professor at the University of Surrey in the U.K. who is leading the group, says: “If I teach my Ph.D. student that and they go on to make a final complex idea, that doesn’t make me an inventor on their patent, so it shouldn’t with a machine.”
U.S. response
The case has drawn business interest partly because patents that list the wrong inventor—or exclude an inventor—could be deemed unenforceable.
That could be a big deal, especially if a company’s AI system discovers, say, a blockbuster drug compound but another party claims it isn’t protectable by patent, says Philip Albert, a partner and patent lawyer with Davis Wright Tremaine LLP, who isn’t involved in this case.
“So we are going to get to a point where it’s going to be really critical to have this question answered,” Mr. Albert says. “We like to resolve these questions before we get into a $100 million dispute.”
In addition to Dr. Abbott and Mr. Thaler, the group behind the case includes four patent lawyers from the U.S., U.K., Germany and Israel. Between October 2018 and August 2019, the team filed patent applications with the U.K.’s Intellectual Property Office, the European Patent Office, the Israel Patent Office and the U.S. Patent and Trademark Office listing Dabus as the inventor.
In August, the U.S. patent office sent a letter stating that the applications wouldn’t be considered unless the applicants listed the inventors involved by their legal names. Later that month, Dr. Abbott and his team filed a response requesting that the office recognize Dabus as the inventor because there was no human inventor. It also asked that Mr. Thaler be granted ownership rights to the inventions.
Dr. Abbott says the applications represent a test case that has implications for fairness, innovation and business certainty. It is unfair, he says, for people who don’t themselves invent to be acknowledged in the same way as people who do. Also, if companies see risks in seeking patents for AI-generated inventions, they may be less inclined to use AI in that manner.
“If Watson and DeepMind become not just competitive with a human inventor, but outperform a human inventor, you would want to be using them in R&D,” he says, referring to AI platforms owned by International Business Machines Corp. and Alphabet Inc., respectively.
“But if [patent offices] are going to say you can’t patent anything that comes out of them, you’re probably not going to get that, or not going to get it as much.”
Parallels to monkey case?
In an interview, U.S. patent office Director Andrei Iancu wouldn’t comment on what’s next for this case but acknowledged this is the first time he has seen an application list an AI system as an inventor. He said his office is working to formulate policy positions on the issue, and that the judicial and legislative branches ultimately will have to decide whether patents can recognize AI systems as inventors.
“Technology is moving very fast, and we need to get ahead of it,” he says.
The U.S. patent office in late August requested public comment on AI, with one objective being to consider “whether new forms of intellectual-property protection are needed.” One of the questions it poses is: “What are the elements of an AI invention?” The deadline for public comment is Nov. 8.
Marc Morley, a patent lawyer with Mintz, Levin, Cohn, Ferris, Glovsky & Popeo PC says he isn’t aware of the details involving Dabus, but he is doubtful it came up with the inventions without any human prompting or provoking. Still, he says, the case raises questions about what it means to invent and who has ownership rights to an AI invention.
“I’m not sure we truly have AI yet where machines are doing the inventing,” he says. “But assuming we will get there in the future, I think ideally we want to rethink things to make sure our laws provide clarity for inventors, owners, investors, so there are not these clouds of uncertainty that are going to slow down innovation.”
Corey Salsberg, vice president and global head of intellectual-property affairs at Novartis AG , says his company is interested in the Dabus case because Novartis uses AI to winnow down chemical compounds that could be used in developing new drugs.
He says the case has some parallels to the “monkey selfie” case earlier this decade, in which an Indonesian macaque took pictures of itself on unattended cameras belonging to photographer David John Slater, who later claimed ownership of the photos. Groups including People for the Ethical Treatment of Animals contested that ownership, saying the copyrights should belong to the monkey.
Ultimately, neither Mr. Slater nor the monkey got those rights, as the U.S. Copyright Office said in 2014 that only photographs taken by humans can be copyrighted, and the U.S. Ninth Circuit Court of Appeals said in a separate decision last year that animals can’t sue for copyright infringement.
Mr. Salsberg’s stance is that regardless of whether an AI system gets credited as the inventor or the person behind it does, the patent should be awarded—unlike in the monkey case.
“What would be a problem is [a patent office saying] you’re not going to get a patent because we believe the AI did the work that a human normally would have done,” he says. “So no one gets the patent. That’s the outcome we’re most concerned about.”
Dr. Abbott says his team ultimately hopes to file patent applications for Dabus in countries including China, India, Japan, Canada, Australia and South Korea.
He says his team is gearing up for a battle that may take five years, with plans to appeal to courts if need be.
“Every country in the world is going to have to decide what they want to do about this,” he says. “AI has shown that it can sometimes improve faster than expected.”
0 comments:
Post a Comment