Keep It Secret Or Safe? AI Developers Face Hard Choice On IP

(February 26, 2024, 5:02 PM GMT) -- Companies developing generative artificial intelligence models will have to decide whether to keep their innovations under wraps as the technology explodes in popularity, experts say — or whether stronger patent protections are worth the risk to publicize their product.

A data analyst using AI technology for Chatbot

Trade secrets are particularly appetizing for technology companies working at the leading edge of artificial intelligence that want to keep the "secret sauce" of their algorithms to themselves. ( Khirisutchalual)

Trade secrets offer protection for innovations that are considered crucial to the success of a business, and these are typically shared among only a few people in the company on a need-to-know basis.

That can range from the secret formulas used by Coca-Cola or KFC to other kinds of "know-how" that are kept under wraps despite the promises that more ironclad patent protections provide. This is particularly appetizing for technology companies working at the leading edge of AI that want to keep the "secret sauce" of their algorithms to themselves, according to lawyers.

Companies have typically turned to patents to protect their innovations. But agencies such as the European Patent Office are limiting the number of patents people can obtain for specific areas.

"AI patent claims have tended to focus on methods of training ... and methods of using [the technology] for a particular purpose — trade secrets' protection is often needed to fill the gaps," Toby Bond, a partner in the intellectual property group at Bird & Bird LLP, said.

The nature of generative AI models also allows the technology that makes the software tick to stay out of the public eye: most of these services are hosted on the cloud rather than downloaded directly on to a computer by the user.

Downloaded software would give customers and competitors some opportunity to peek under the hood, while cloud-based programs create further barriers to those who want to reverse-engineer the technology.

"This makes it harder for competitors to discover how your technology works, allowing companies to rely more on trade secrets to protect their technology," Bond said.

This comes alongside a general push to solidify trade secret protection from regulators. To this end, the U.K. government introduced the Trade Secrets (Enforcement, Etc.) Regulations 2018 to solidify the British position on IP.

The government has also stated that there is a "growing emphasis" on the role of trade secrets in the plan to boost innovation, and has given assurances that it wants to position the U.K. as a leading force in AI development.

Regulators have also acknowledged the role that trade secrets could play in protecting AI technology directly. The government opened a consultation into the matter in March 2021, and is still taking responses.

The European Union and the U.S. both introduced their own regulations to harmonize protections and protect against the misappropriation of trade secrets at the end of the 2010s, in an "alignment of the stars" for IP, according to Gaëlle Bourout, a counsel at Linklaters LLP.

Technology trade secret litigation surged in the U.S. following this legislation, and lawyers say that there might be a similar increase in the EU and U.K., potentially giving companies grounds to better enforce their confidential IP.

"History shows that Europe can follow up on the U.S. with a certain lag," Bourout said.

But companies must ensure that they keep an internal record proving that the information is confidential. And they have to be able to prove that the information is limited to a "need-to-know" basis for selected employees.

A company could have a tough time proving any potential infringement of its IP in court or arbitration if it does not have a proven track record that it has treated the information as confidential.

"Even within an organization you can lose the necessary quality to protect something as a trade secret … it needs to be on a 'need-to-know' basis," Sarah Mountain, a partner at Reynolds Porter Chamberlain LLP, said.

Detecting the infringement itself can be a tall order. A company that suspects a competitor is using its confidential technology can find it tricky to prove that it has incorporated this "know-how" into developing its own products. This is particularly true when it is information that is meant to be kept under wraps.

"If your innovation is something that your competitor could discover from taking your products and service, then patent protection is more likely to be appropriate," Bond of Bird & Bird said.

--Editing by Joe Millis.

For a reprint of this article, please contact

Hello! I'm Law360's automated support bot.

How can I help you today?

For example, you can type:
  • I forgot my password
  • I took a free trial but didn't get a verification email
  • How do I sign up for a newsletter?
Ask a question!