Mark Cuban: AI Agents Lack Human Judgment in Business Decisions

Mark Cuban has argued that companies using AI agents as a replacement for employees could overlook human judgment in business decisions. The billionaire investor and Shark Tank judge said that human workers understand the consequences of their actions in ways AI models currently do not. His comments came in response to a discussion on the microblogging site X (formerly Twitter) on whether AI has begun affecting employment in the US Replying to Box CEO Aaron Levie on X, Cuban wrote that businesses should not treat AI and white-collar workers as mutually exclusive. "People know what will get them fired," he said, adding that AI models lack real-time judgment and awareness of workplace dynamics.

In his X post, Cuban wrote, “Models don’t know the consequences of their actions. People know what will get them fired. Models suffer from information latency. People understand what they see right in front of them. 2 skills AI won’t have for a long, long, long time. If ever. 2 skills that are invaluable to every business decision. How often in business do you need to “read the room” and make a decision? Add AI productivity to the real-time capacity and judgement of humans, and you will get the greatest return on both investments.The challenge is management understanding how to leverage the combination, and forgetting the original expectation that AI and white-collar workers are mutually exclusive. They are not. Done right, the combo is a business propellant and competitive advantage."

Debate began with Anthropic's report on AI and employment

Cuban's remarks followed a post by Anthropic's Head of Economics, Peter McCrory, who shared a report titled “Why hasn’t AI increased unemployment?” The report argued that despite widespread AI adoption in the US, there is currently no evidence that AI has caused a material increase in unemployment. McCrory said the US labour market remains close to full employment, while AI has so far functioned as a "skill-biased" and labour-augmenting technology that complements human expertise rather than replacing it. He noted that although AI automates some tasks, human oversight remains necessary for planning, evaluation and complex decision-making. The report also said workers in AI-exposed occupations have expressed greater concern about job losses, but unemployment rates in those roles have not increased unexpectedly. It added that AI capabilities are advancing rapidly, though whether future systems could lead to broader labour displacement remains uncertain.

Box CEO Aaron Levie says AI still depends on people

Responding to McCrory's findings, Levie, the chief executive of AI-based content management tool company, said current data suggests AI has had a smaller impact on employment than many expected because it still depends on people to generate value."Very good post from the Head of Economics at Anthropic. They’re finding that jobs have been less negatively impacted by AI than expected, as we continue to see time and time again in the data,” he wrote. Levie added that AI is currently better suited to automating specific tasks rather than entire jobs, allowing workers to increase productivity instead of being replaced."The reason for this is that AI - at least so far - still requires people to operate to produce value in most cases. Most jobs can’t be fully automated with AI, only certain tasks in those jobs. And when you automate specific tasks, you actually can get even more output from those jobs, raising the demand (or at least maintaining it) in many cases,” he added. Quoting McCrory's report, Levie wrote: “So far, AI is both skill-biased and labour-augmenting. It complements domain expertise. It relies on humans in the loop to direct and evaluate the most complex work. And it rewards AI proficiency. Model capabilities are improving fast, but remain stubbornly jagged. To fill in the pockets of the jagged frontier, expert oversight is needed to steer incredibly capable AI systems, and to recover when they falter.” He further argued that this pattern is already visible in software engineering, where AI coding agents have increased output while continuing to require developers to manage and supervise their work."But there will be plenty of other domains of work where demand remains strong in a world where agents can accelerate the output of that job. Jevons paradox is alive and well,” Levie highlighted.