This week, OpenAI quietly published a statement on its website pledging not to use its patents offensively.
Asserting its commitment to the principles of “broad access” and “collaboration,” OpenAI said that it would only use its patents in a way that supports innovation. “We pledge to only use our patents defensively, so long as a party does not threaten or assert a claim, initiate a proceeding, help someone else in such activities against us, or engage in activities that harm us or our users,” wrote the company.
But experts say the claims are little more than fluff.
In contrast to patent pledges like IBM’s 2005 promise not to use 500 specific patents against open source software makers, OpenAI’s is ambiguous and vague, according to MBHB partner Mike Borella. It’s not clear what “defensively” means, exactly — or which activities OpenAI might perceive as “harming” the company or its users.
“The last phrase seems to be the exception that swallows the rule,” Borella told TechCrunch. “It can be read in so many different ways, including to cover all of OpenAI’s competitors in the marketplace and parties who fairly criticize ChatGPT’s deficiencies.”
OpenAI also doesn’t have much of a patent portfolio it could assert if it wanted to, he argued; its IP strategy is much more dependent on trade secrets — things like confidential training data and methods. Patent applications with the U.S. Patent and Trademark Office aren’t generally published until 18 months after filing, but, at present, only a handful of OpenAI’s have been granted.
Then there’s the fact that OpenAI’s patent pledge is just that: a pledge. Voluntary, not binding, like a contract or warranty.
It’s not clear what, if any, legal weight it might have, says Shubha Ghosh, a Syracuse University professor specializing in IP law. In any case, it won’t limit OpenAI from enforcing its rights under copyright, trade secret, or contract laws, he noted.
“Computer source code that guides ChatGPT would be protected by copyright law and trade secret law,” Ghosh said. “Terms that OpenAI includes in its terms of service would still be enforceable as contracts. Should OpenAI engage in conduct that deviates from the pledge, for example by suing another company or an individual for patent infringement, what OpenAI has lost temporarily is credibility.”
Borella called OpenAI’s statement “public relations virtue-signaling”: an attempt to curry favor with the tech community and regulators, along the lines of Tesla’s ill-defined patent pledge in 2014.
“The statement as a whole ends up being nothing more than a paper tiger,” Borella said, “rather than a substantive attempt to foster healthy competition in the marketplace.”
