When Graciela Dela Torre developed chronic pain while on the job at a freight company in the Chicago area, she did what millions of Americans have done: She filed a disability claim.
After a couple of years, her company’s insurance carrier decided she no longer qualified for compensation. Again, Dela Torre did what millions of Americans have done: She hired a lawyer, who negotiated a settlement payment for her. Case closed.
But Dela Torre had second thoughts, and she did what, yes, millions of Americans have done: She consulted ChatGPT, the online artificial intelligence tool that enables everyday people to “chat” with an AI system.
Dela Torre didn’t like her lawyer’s flat response that nothing could be done to reopen her case, which had been permanently dismissed as part of the settlement. Had her lawyers “gaslighted” her? she asked ChatGPT.
Yes, the system replied, and it didn’t stop there. At her direction, ChatGPT performed extensive legal research and drafted piles of legal documents. That enabled her, with no lawyer, to launch a multipronged effort to set aside the binding settlement and revive her claim.
Today, Dela Torre’s AI-fueled legal odyssey has turned into a closely watched case in Chicago’s federal courts and beyond.
Dela Torre’s target, Nippon Life Insurance Co. of America, has sued OpenAI, the company behind ChatGPT, alleging its electronic platform was practicing law without a license. Both sides have lawyered up with real, live attorneys from big-time law firms, no doubt billing four figures an hour as they fight about the AI version of themselves.
The insurance company has alleged that a rogue AI robot masterminded Dela Torre’s campaign to improperly challenge her settlement agreement, generating a flurry of baseless and abusive filings that cost $300,000 to contest. OpenAI has fired back that ChatGPT is not a lawyer and doesn’t claim to be one, and when she consented to its user agreement, Dela Torre agreed not to use it as a substitute for professional advice.
OpenAI wants the case dismissed and a hearing is scheduled for June 30 in Chicago. But even if this case goes away, rest assured the issues it raises are just beginning to be heard.
ChatGPT passed the Uniform Bar Exam with a combined score of 297, far above average, so in one important way, at least, it’s “smarter” than the typical novice lawyer. The research and documents it produced for Dela Torre sure look authentic. But at least some of it was gobbledygook. And based on the docket, the federal judge who handled her case grew exasperated with the avalanche of filings. Still, they had to be laboriously contested.
As in practically every white-collar occupation, lawyers are embracing AI to make their practices more efficient and effective, and many lawyers use ChatGPT to do some types of legal work. Yet millions of legal cases in the U.S. involve non-lawyers, or so-called pro se litigants. Does anyone doubt that in many of those cases, AI already is being used? Why should lawyers be able to use AI and not everyday people involved in legal proceedings?
For one, ChatGPT is not a member of the bar, bound by its ethics, or admitted to practice before the federal courts. So, if it is indeed practicing law, it is breaking the rules. On top of that, AI models have been known to cite nonexistent legal cases, a practice nicknamed “hallucinating” that gets a lot of attention when human lawyers cite made-up precedents in their AI-enhanced pleadings.
Most important to us, AI lacks human judgment. We don’t know the details of Dela Torre’s relationship with her human lawyers. But advising a client to forget about trying to reopen a duly closed case strikes us as sensible advice, unless something drastic has changed, whereas AI models tend to tell people what they want to hear so they’ll keep using them.
AI has become one of the most important issues of the day. Governments around the world are struggling to control it. Militaries are weaponizing it. Companies are spending fortunes adopting it. Political parties are taking sides on it, knowing that wealthy donors love AI even as many voters increasingly hate it.
Successful AI practitioners like OpenAI are poised to raise trillions in initial public offerings in the months and years ahead, expanding their influence. AI spending continues to power economic growth and, depending on the day, the stock market. At the same time, we hear rumblings that AI has become way too costly, and that it’s not remotely paying off.
We don’t know where today’s AI fever will lead, but as for the case between Nippon Insurance and OpenAI, we asked ChatGPT who’s going to win it.
“It’s impossible to predict,” the platform told us, but also helpfully offered to perform a risk analysis to give “a clearer picture of who might have the edge.” Its risk assessment showed OpenAI with a 57.5% chance of prevailing in the lawsuit, while Nippon had a mere 42.5% chance.
Surprise! ChatGPT predicted that ChatGPT had the edge.
So, is that another case closed? Don’t count on it. No case will ever be closed if machines are permitted to do our thinking for us.
Submit a letter, of no more than 400 words, to the editor here or email letters@chicagotribune.com.
