Experts flag 3 areas where AI does more harm than good

Companies rushing to deploy AI across every business function risk alienating customers, eroding brand trust and violating regulatory norms, practitioners warn.

The most skilled AI practitioners are defined by what they leave to humans, not what they automate. Empathy in customer service, authenticity in marketing and transparency in regulated fields are three areas where AI does more harm than good, industry practitioners say.

"When someone calls in seeking to cancel an account, it's a cinch for AI to handle that. But let's say it's a widow calling up to cancel her husband's account — do you really want AI handling that interaction?" said Dan Leiva, a former executive at Apple and eBay who pioneered AI for customer service.

Leiva's concern extends beyond empathy. As head of payment operations at eBay, he said one of his biggest worries about generative AI was whether companies could explain decisions the systems made after the fact. This transparency problem is acute in regulated industries including payments, medicine and law, where large language models remain biased black boxes that are capricious in their decision-making, he added.

Companies that over-automate risk losing institutional knowledge and talent pipelines, Leiva said. "We will see who over-automated, and which ones lost the knowledge that lets them differentiate themselves from other companies," he added. The warning echoes Project Plowshares, the US government's 1960s effort to use nuclear bombs for peaceful earthmoving — a program that ultimately demonstrated not new uses for atomic weapons but when they should never be used at all.

The Empathy Gap

Natalie Desseyn, a double-board-certified nurse practitioner in family and psychiatric medicine, uses six AI tools for everything from taking notes during sessions to filing insurance claims. The tools save her 15 to 20 hours a week, allowing her to handle 300 patients as a solo practitioner. Yet the most important function these tools serve, she said, is making it possible for her to be fully present during psychiatric counseling sessions.

"In psychiatry, if you're not engaging with the patient, you're going to miss a lot of symptoms," Desseyn said. "I have a couple of patients that I'll ask, 'Are you thinking about hurting yourself?' And if I'm not looking at them, I'm not going to get the right answer."

Authenticity and Transparency Under Pressure

In marketing, the premium has shifted to human storytelling as AI-generated content floods social platforms. Big companies are now seeking human writers to break through the AI slop on LinkedIn and elsewhere, according to Christopher Mims, the WSJ columnist who authored the analysis. The rise of short-form video influencers — whose direct-address content is harder to fabricate with AI — reflects the same trend.

The problem extends to the broader information ecosystem. A research study titled "The Ghost Couple" found that large language models default to a small set of high-probability names — such as Elena Vasquez and Marcus Chen — when generating fictional experts, creating a feedback loop where AI-generated content pollutes the web and trains future models on synthetic data. "The web is an unintentional archive of LLM behavioral fingerprints," the researchers wrote.

Mike Walsh, CEO of LexisNexis, said generative AI has shifted how customers use the legal research service — they now spend more time drafting documents than researching case law. The reason they can do so with confidence, he said, is that LexisNexis's system is only allowed to cite cases it can verify exist in its databases, unlike generic chatbots that hallucinate nonexistent cases. Lawyers who use such systems still review every AI-generated output, Walsh emphasized.

For investors, the implications cut both ways. Companies that deploy AI without maintaining human oversight risk regulatory penalties, brand damage and customer churn that could outweigh efficiency gains. The gap between what CEOs expect from AI-driven productivity and what employees report is already wide. Firms that replace junior workers with AI risk destroying their own talent pipelines — when senior engineers leave, no humans remain to review the AI's work. The winners in the AI era will be companies that use the technology to augment human judgment, not replace it.

This article is for informational purposes only and does not constitute investment advice.