This week, OpenAI silently released a statement on its website promising not to use its patents aggressively. Emphasizing its dedication to the principles of “broad access” and “collaboration,” OpenAI stated that it would only use its patents in a manner that promotes innovation. The company affirmed, “We commit to using our patents defensively, as long as a party does not pose a threat or make a claim, initiate legal action, assist others in such activities against us, or engage in actions that harm us or our users.”
However, experts argue that these claims are mostly insignificant. Unlike patent commitments like IBM’s 2005 pledge not to use 500 specific patents against open-source software developers, OpenAI’s promise is unclear and vague, according to MBHB partner Mike Borella. The meaning of “defensively” is ambiguous, as is what activities OpenAI might consider as “harmful” to the company or its users. Borella expressed, “The last phrase appears to be the exception that nullifies the rule. It can be interpreted in various ways, even to encompass all of OpenAI’s competitors in the market and individuals who criticize ChatGPT’s shortcomings.”
Borella also noted that OpenAI lacks a substantial patent portfolio to utilize if necessary; its intellectual property strategy relies more on trade secrets, such as confidential training data and methods. Although patent applications with the U.S. Patent and Trademark Office are not typically published until 18 months after filing, only a few of OpenAI’s applications have been approved so far. Additionally, OpenAI’s patent pledge is merely voluntary and not legally binding like a contract or warranty.
Shubha Ghosh, a Syracuse University professor specializing in IP law, mentioned that the legal weight of OpenAI’s pledge is uncertain. Regardless, it will not prevent OpenAI from enforcing its rights under copyright, trade secret, or contract laws. Ghosh explained, “Computer source code that guides ChatGPT would be protected by copyright and trade secret laws. The terms included in OpenAI’s service agreements would still be enforceable as contracts. If OpenAI were to violate the pledge, such as by suing another company or individual for patent infringement, the company would temporarily lose credibility.”
Borella criticized OpenAI’s statement as “public relations virtue-signaling,” suggesting it is an effort to gain favor with the technology community and regulators, similar to Tesla’s vaguely defined patent pledge in 2014. He stated, “The statement ultimately amounts to nothing more than a hollow promise, rather than a genuine attempt to promote healthy competition in the market.”
