Mark Cuban believes businesses need to embrace AI to succeed, but he says those that use it incorrectly are doomed to failure.
“There’s going to be two types of companies: those who are great at AI, and everybody else,” the celebrity billionaire of “Shark Tank” fame said. “And the ‘everybody else’ is going to fail because AI is such a transformative tool.”
The comments came during a call with Clipbook founder Adam Joseph, whose startup secured a seven-figure investment from Cuban. Business Insider reviewed a recording of the call.
On the call, Cuban explained to Joseph how he believed everyone, from entrepreneurs to employees, should — and shouldn’t — use AI.
Like fellow shark Kevin O’Leary, Cuban thinks AI will have a substantial positive impact on businesses that implement it well. However, business leaders need to understand the intricacies and distinctions between different AI tools and not treat them as interchangeable. He said using AI ineffectively could turn helpful tools into an expensive distraction.
“Because AI is continuously changing, you need to just have people — and, really, every CEO — taking the time to understand every nuance of every new tool that comes out,” Cuban said.
The AI revolution isn’t going anywhere
The world is still “in the first inning of the first preseason game” of the AI revolution, Cuban said, even though generative AI tools like ChatGPT have been out for more than three years, while other forms of AI, like machine learning, have been around for decades.
Tech companies like OpenAI, Google, Microsoft, Meta, and Elon Musk’s xAI are spending tens of billions of dollars to win the AI wars.
Cuban said that “it’s too early to tell” which of those companies will succeed, or if someone else will create the go-to AI chatbot.
Craig Sjodin / Getty
“They all want to be the destination that everybody turns to, but it’s not that straightforward, and we don’t have a winner yet,” Cuban said.
Businesses that discount the power of AI are destined to get disrupted, in Cuban’s view.
“If I’m going to compete in an AI world, data or information is more valuable than gold, more valuable than oil,” Cuban said.
But for all the hype about AI, Cuban is clear-eyed about the technology’s limitations. AI tools can be mistake-prone yet hyperconfident, and chatbots aren’t always smart.
“AI is stupid,” Cuban said. “But it’s somebody who’s a savant that remembers everything.”
Cuban likened AI chatbots to people who have minds like a steel trap. These tools are able to instantly recall and process tons of information, then aggregate it in one place.
“It does a really good job of assembling all those things that it collected and presenting that just as somebody who has a great memory,” Cuban said.
AI isn’t a cure-all
AI chatbots have holes besides so-called hallucinations, Cuban said. AI tools sometimes don’t pull up-to-date information. They can also be unclear about how they reach their conclusions, as their algorithms are opaque and can cite faulty or inaccurate links.
He also said that people often presume the “AI models they’re using or creating” will provide all the answers they need, but that’s “just not the case.”
Cuban said AI can actively harm businesses that use it incorrectly or those that don’t understand its capabilities.
Employees who use standard versions of tools like ChatGPT could be compromising sensitive company information. Similarly, businesses that post their work online must also realize that they could be giving it away for free to web-scraping chatbots hungry for new information.
“Companies are learning now that their IP is incredibly valuable,” Cuban said. “Two years ago, last year, two months ago, they might have just posted everything on the net to show how smart they are, or shared everything in a proposal to show how smart they are. Now, you have to be really careful with your IP.”
Cuban said academics or hospital researchers must pivot in the AI age away from a “publish or perish” mindset, where they share their findings widely in peer-reviewed journals.
“Now, doing that’s the biggest mistake you can make, because all you’re doing is training somebody else’s models,” Cuban said. “And so you’ve got to be able to understand what IP you need to be able to protect, how you’re going to disseminate that IP, whether or not you want to sell it, or keep it for your own models, and how you acquire information.”
