After finally being issued an artificial intelligence technology patent, the congratulatory feeling can quickly lead to apprehension that the secret sauce powering the company's technology is exposed to all. Concerns over tech details and AI's ongoing development can make filing patents too expensive and time-consuming for many contract management companies.

Seal Software co-founder and chief technology officer Kevin Gidney called it the "double-edged sword" tech companies face when they file a patent for AI technology.

"We stopped doing patents for a five-year period because the market was becoming extremely overcrowded with what we were doing," he said.

Still, Gidney noted there are positives to a patent when it's granted for a specific function.

"[If] you can get the claims to be so pointed with no other way to do that particular function, then it's a powerful patent," he said. "That's very rare to find those types of things to be frank, but it is possible."

AI-assisted contract reviewer startup BlackBoiler, however, thinks it has four patents that are distinguishable. 

"Most AI applications are not coming from the legal sector, and what's unique to our patent is we haven't patented AI but a process that can leverage AI," said BlackBoiler founder and CEO Dan Broderick. 

BlackBoiler was issued four patents from the U.S. Patent and Trademark Office, including a method and system for suggesting revisions to an electronic document and a patent for systems, methods and computer program products for a clause library.

Broderick noted the patents covering a method and a process can be an effort to define the scope of patent protection for a specific technology such as AI.

He agreed it's unusual for startups to file a patent application and invest significant time and money into a process early in a company's formation. But after noting the estimated $35 billion spent on high-volume contract review by Global 2000 and Fortune 1000 companies, Broderick said his company's patents address "a huge pain point we are tapping, and I know this is a problem that is ripe for significant automation."

He added that "as a startup, we designed an intellectual property and trade secret strategy that will allow us to deliver on the efficiency promises to our clients while giving us protection from organizations with larger marketing budgets that may muddle the market with lesser technology."

Jerry Ting, CEO of AI-backed automated contract management platform Evisort, said his startup also considered filing patents. However, when factoring how quickly the technology is changing in the sector, they decided "it wouldn't mean much in two to three years."

Plus, as soon as the patent is listed on the USPTO and other patent regulators' databases, the details of the patents are revealed and can be easily duplicated with slight modifications to avoid patent infringement, he added.

Indeed, BlackBoiler's Broderick said AI patents publicly disclose an entity's technology, though he noted that is in exchange for a "limited monopoly as defined by the patent claims." Still, he added that entities can also protect aspects of their AI tech as trade secrets, if that technology cannot be easily copied or reverse engineered.

However, such protection might not happen often. Gidney and Ting point to tech giants Google, Microsoft and Amazon creating and sharing AI algorithms among peers and smaller tech competitors as the unspoken rule that patenting these algorithms stifles innovation.

"The approach is that certain things you can't patent and what you really shouldn't patent as a good member of society. We all build our software on open-source libraries and algorithms that are made available to us," Gidney said.

To be sure, Gidney doesn't foresee many contract analysis companies filing AI tech patents.

"I don't see a mad rush until an extremely large player emerges in the industry," he said, "Then you would force some of the new entrants out of the market."