This week’s Senate hearing featuring OpenAI’s Sam Altman raised a multitude of questions about how to regulate the technology to protect voters from deep fakes or half-truths, but also how to protect intellectual property, including even the AI algorithms being written by companies across multiple industries.
One underlying question on the minds of many, including senators themselves, was whether government and the courts are up to the task of regulating AI and adjudicating cases when conflicts arise. Many artists and writers of software and songs worry AI will steal their creations, especially for the training of AI.
“I do think we are up to the task” of dealing with AI regulations and adjudication in the nation’s courts, said Elizabeth Manno, a partner at Venable LLP who specializes in IP. “Our system is designed to adapt to change, but there are growing pains right now. I’m heartened the Senate is taking it seriously. We do need to get ahead of AI.”
Manno views the rapid adoption of ChatGPT and generative AI with a mix of concern and excitement, she said in an interview with Fierce Electronics. “I’m concerned as a citizen about the accuracy of what’s out there. I’m concerned but also excited. It’s amazing to see the technology develop. However, the aspect of fake news stories and deep fakes is very concerning.”
At the same time, there are already programs using AI available to help companies and attorneys conduct such tasks as researching prior art when beginning to file patent applications, she noted. Questions remain about “how much more efficient we can get with this new invention.”
As an attorney specializing on IP, she said, “I don’t want people to be discouraged about getting IP protection. It might become a little more of an in-depth process. People might have to think more creatively about having AI in the picture.”
Tips for creating patents, copyrights with AI
Regarding IP protections with AI, she offers tips and strategies to her clients and companies at large. “The law is running a little bit behind the tech, but there are strategies to protect IP and you have to get ahead…Any company needs a cohesive IP strategy, AI or not.”
A big concern for companies will be in differentiating what was created by humans and what was created by AI. “That insight informs what types of IP protection you need,” she said.
So far, the law holds that AI cannot be an inventor of a patent. “To get a patent, you have to think about what was the human element or human involvement. However, using AI with human involvement doesn’t mean you can’t get a patent.” For now, an AI cannot be the inventor “and that’s mostly true around the world.”
Also, AI cannot be an author when seeking a copyright, Manno said. “If you have something you want to protect, decide what was the human involvement and then you can copyright [that part],” she said.
In any event, consult an IP lawyer on the specifics to see if a patent or copyright is possible, she advised.
One way to protect content or a product or service is to develop a trade secret internal process so that something a company wants protected will only be accessible by a few insiders. “The benefit is you don’t have to get approval for authorship with a patent or copyright,” Manno said. This approach means if a trade secret needs to be shared with a customer or partner, it should be protected with a signed confidentiality agreement.
“A trade secret has to be something valuable to a company, of specific value, and you can still get IP protection,” Manno explained. “If you create your own AI algorithm, you can still protect those things” subject to patent and copyright restrictions. “You can still get patents on your own inventions.”
Manno noted that many companies, from startups to enterprises, are using AI to generate code where the company still has an interest in protecting its own code. Admittedly, she said, “there is still an ongoing struggle on what can be protected.”
Some news operations and online publications, as well as a broad array of companies in other industries, are already publishing art or other content created with AI.
“You have to be cautious,” she warned. “You want to make sure it is not content protected. That’s not to say you can’t use it, but it’s worth looking into,” preferably with the help of an attorney trained in the matter. As with the issue of deep fake animations that rely on using a living human’s face, the question arises about what is the line between permitted use or not.
“It’s the wild west right now,” Manno said.
RELATED: Google, IBM unveil new AI plans as competitive tension increases