Lawyers warn of serious risks

A version of this article first appeared in CNBC’s Inside Wealth newsletter with Robert Frank, a weekly guide for the high-net-worth investor and consumer. subscription To receive future issues, directly to your inbox.

Attorney Tasha Dickinson said she gets calls every week from clients asking about legal advice they got from ChatGPT, Claude or another AI-powered chatbot. She said that some do not admit it, but she can tell by the way they are questioned.

One client, a high-net-worth Florida resident, asked Dickinson about setting up a community real estate trust — an attractive option for married couples — saying he got an AI suggestion to save taxes for his heirs, she said. Dickinson quickly pointed out a problem: the client’s wife had recently died.

“I said, ‘Well, you understand that a community property trust is between husband and wife, right?’ There was silence on the phone,” said Dickinson, a partner at Day Pitney. “They’re saying, ‘Oh, well, the AI ​​thought it was a good strategy.’ Well, in universe, that might be a good strategy, but it’s not a good strategy for you.”

Wealthy lawyers tell Inside Wealth that their clients are increasingly using AI not only to research tax topics but to second-guess their lawyers’ advice. While some lawyers say AI helps clients ask informed questions and learn basic concepts, they also say it poses headaches and legal risks.

Robert Strauss, a partner at Weinstock Manion, said many clients uploaded trust documents to AI systems and came back with a list of questions and suggested amendments, forcing Strauss to defend his work and explain why the AI ​​recommendations didn’t fit the client’s situation.

“Questions are good, but they result in us spending more time on the issue than we normally would,” he said. “We ended up spending two, three, four hours of time dealing with things that didn’t amount to anything yet. And I didn’t actually receive any actionable suggestion from that process.”

The result, he said, is that the client does not trust his lawyer.

More troubling, Strauss said, is that customers are sharing sensitive information with large-scale forms, raising data privacy concerns and legal pitfalls. Strauss said his firm is currently reviewing its client contract to warn clients that using AI-powered chatbots like this could void attorney-client privilege.

In February, a federal judge ruled that a criminal defendant’s conversation with Claude about his legal defense strategy was not protected by attorney-client privilege.

“What keeps me up at night about AI? It’s not that AI is wrong sometimes, because I can correct those mistakes. It’s not that people review my work on AI, because I have a lot of confidence in my work,” Dickinson said. “My concern is that when people put documents and do these searches in AI, they are waiving attorney-client privilege, and that’s a big problem.”

Asking a chatbot how to protect your assets with a prenuptial agreement or how to sell your business while paying less in taxes, for example, could be used against you in court, warned Dan Griffiths, director of wealth strategy at Huntington Bank.

While high-net-worth clients generally have access to and afford the best legal advice, they, like all of us, enjoy the convenience of AI, according to Griffiths.

Dickinson added that cost savings remain an attraction. “Paying for professional services is not fun,” she said. She added that many of her clients are confident entrepreneurs.

“A lot of our clients have been very successful. I mean, they’re smart, right? And they have a drive for knowledge,” she said. “I think some people make the mistake of assuming they understand this more than they actually do.”

Using these AI tools “gives a false sense of knowledge,” she said.

In some ways, this is not a new problem. Clients often give their attorney suggestions they got from a friend at the country club or from an article. Dickinson described it as “a more sophisticated form of cocktail party talk.”

The trend is not one-sided. Many lawyers use artificial intelligence in their professional and personal lives. This has led to major headline-making errors such as feeds with fake citations.

But lawyers say few clients are knowledgeable enough about AI and the law to write an effective claim.

Ed Renn of Withers gave the example of a client who wanted to transfer unlimited assets to his wife on the advice of ChatGPT. However, the client did not mention that his wife was foreign-born, meaning he could not take advantage of the unlimited marital deduction without a special type of trust, according to Ren.

“If you don’t know exactly what you’re doing, it’s garbage in, garbage out,” he said.

Ren added that AI tools appear to make more errors on more complex topics such as international taxes and are not up to date with new legislation or guidance issued by the IRS.

Determining how to transfer your wealth to your loved ones requires a more complex discussion than ChatGPT is prepared for, Griffith said. He said there are rarely easy answers when deciding, for example, how to divide assets between children from the first marriage and the second.

“If your client asks you: ‘Hey, if I make this trust, will my son have access to the money I give him at some point?’ The answer shouldn’t be ‘yes’ or ‘no.’ The answer should be: ‘Tell me more about your relationship with your son, or what it’s like?’” AI tends to be solution-oriented and try to find some way to get to the yes answer. He doesn’t do a good enough job of saying, “You know what? Let’s get to the heart of your question.”