AI's legal implications run deep

Using artificial intelligence in work has myriad legal implications and consequences. What those are, exactly, remains a work in progress given how new the technology is, according to a leading intellectual property attorney. 

“It seems like the legal system has finally got its collective head around the internet generally and now (there’s) internet 2.0, the metaverse and AI," says Mark Nieds, chair of the intellectual property group at Fort Myers-based law firm Henderson Franklin Starnes & Holt. "Technology is outpacing the speed of legal adoption.”

But Nieds is already seeing the risks of AI get real.

Recently, an attorney in New York, for example, used AI to write a brief for an upcoming court case. But the AI’s brief cited nonexistent cases. Now that case is tackling a whole other slew of issues, Nieds says, including the rights of each party and malpractice exposure of the attorney. “It’s one of those situations where AI sounds like a good idea but really can backfire,” he adds. 

The biggest legal issues around AI, Nieds says, are twofold: The output is not verified, and there’s potential copyright infringements. Other concerns, says Nieds, include:

• Some states are starting to regulate AI under general data privacy laws and employment law, particularly on requiring employers to disclose if they use AI in a job interview. Nieds notes there’s concern it could lead to discrimination due to biases built into AI, as well as the AI platform making assumptions based on how a question is answered. 
• According to the law, machines can’t own copyright or patents. So if you were able to get AI to produce an invention for your company, it likely wouldn’t be patented. How do you protect it? 
• Disclosures. All companies have trade secrets. If you use AI and you’re inputting your trade secrets to the AI, does that count as disclosure? Consider, says Nieds, the famous Coca-Cola recipe. If Coca-Cola wants to make the recipe better using AI and inputs all of the ingredients and how it’s currently made, AI could make a suggestion. But then the question is, will that recipe no longer be a secret? 
• Responsibility. If an AI platform is derived to drive a car that hits and kills a pedestrian, where is the liability? 

“I’m not really worried about it in the Terminator sense, but more so advising clients to go slow,” Nieds says of AI’s presence. “Embrace it, learn about it and see if you can utilize it, but look at it closely before diving headfirst.”