Does AI have consistency issues?
That’s the question posed by billionaire entrepreneur Mark Cuban in a post on X Monday. He said he is “coming to the conclusion” that the “biggest challenge” facing both enterprise AI and consumer tools is simple: users don’t get the same answer when asked the same question twice.
“It remains impossible to ensure that everyone gets the same answer to the same question every time,” he wrote. “AI doesn’t know the consequences of its output.”
Cuban’s points highlight important design differences between AI systems and traditional software.
Most enterprise tools have historically been built on deterministic rules. That is, it ensures that the same input produces the same output for the user.
Generative AI systems operate based on probabilistic structures. Popular current models such as OpenAI’s GPT-5.5, Anthropic’s Opus 4.7, and Google’s Gemini 3.1 generate responses based on probabilities and choose from a range of possible responses, rather than following a single fixed path.
In practice, this means that the same question can generate different responses, making AI systems less predictable in business environments. It can also cause AI hallucinations.
Some X users pushed back against Cuban’s points, arguing that this volatility is part of the trade-off. They say that for more open-ended or creative tasks, multiple answers may be valid, and enforcing strict consistency may limit the tool’s usefulness.
Still, in his original post, Cuban cited the variability of AI as a reason why human judgment is becoming more important.
“Judgment and the ability to counter AI output are becoming increasingly necessary and valuable,” he wrote. “This increases the value of your domain knowledge by the second.”
A focus on how people use and question AI is something Cuba frequently emphasizes.
For example, speaking Wednesday on the Big Tech Podcast at the Dallas Regional Chamber’s Convergence AI event, Cuban described the growing disparity in how workers use technology.
“I think we’re now divided into two types of ways to use AI, or two types of people: people who use AI so that it doesn’t have to learn anything, and people who use AI so that it can learn everything,” he said.
He added that those who treat AI like a “drunk intern” risk falling behind, while those who use it to develop skills will benefit.
Cuban also shared practical ways for people to engage with technology. He recently spoke to Business Insider about three starter questions to ask Anthropic’s Claude to help you launch your small business.