SharkNinja CIO Velia Carboni: Let AI go 'a little bit rogue'

Carboni's remarks came during a panel at Fortune's Most Powerful Women Summit in Santa Barbara, Calif., where executives pushed back on recent AI doom headlines. Much of that alarm, the panel said, stems from AI labs' own stress tests — in which frontier models are pushed to deceive, break rules, or resist shutdowns — and from disclosures that have at times been tardy or less-than-fulsome.

Carboni's position goes further than her peers on the panel by defending the labs' boundary-pushing as a strategic necessity. At SharkNinja, a $27 billion company launching dozens of new products per year in up to 15 countries, she said guardrails are firmly in place, but she doesn't lead with policing when talking to employees because "I don't want to limit creativity." The company is also digitizing some of the judgement of CEO Mark Barrocas and his top commercial leaders to make decisions such as where to move scarce inventory — an effort Carboni described as sparing people from having to "pick up the phone and call Mark."

The panelists converged on judgement as the line between what AI can be trusted to do and what still needs human care, a theme Carboni's approach illustrates: experimentation at the lab level paired with guardrails and human judgement inside the business.