Ethan Mollick’s 4 guiding principles for leading with AI

Take a survey of leaders about AI in the workplace, and you’ll run the gamut of opinions. On one end of the spectrum, the AI evangelists argue the technology has ushered in a golden age of creativity, cut costs, and delivered productivity gains. On the other, the naysayers fret that AI overcomplicates workflows, opens companies to security and liability risks, and degrades the human element that makes work worthwhile.

But we’d wager most leaders would fall somewhere in the middle. They realize that AI is taking over tasks once done by humans and worry that they’ll fall behind without it. They also recognize that AI is far from perfect and may fail at inopportune moments. For instance, the tech website CNET once used an AI to write several articles wound up containing many boneheaded errors. The fallout was a journalistic and PR disaster for the company.

Let that settle in: Leaders at a technology-focused media company couldn’t properly deploy AI. What chance do the rest of us have at getting it right?

The answer is better than you’d think. Ethan Mollick, a professor at the Wharton School, specializes in entrepreneurship and innovation. His research has led him to become a prominent voice in the AI field. Through his research, he has discovered a framework to help business leaders navigate AI’s integration. It consists of four principles that offer a roadmap to bring AI in without tossing human judgment out.

During an interview with Big Think+, Mollick shared these four guiding principles alongside some advice for making them work. Let’s dive in!

#1. Always invite AI to the table

Mollick’s first recommendation: Don’t shy away from AI. You need to embrace it because that’s the only way to understand where and how it can be applied effectively. Use it to critique an idea, write memos, record meeting minutes, give feedback, summarize reports, or produce content.

“Use AI for everything you legally and ethically can,” Mollick says. “You do not know what AI can do, and likely, the creators of the AI system you are using do not know what it can do for you and your industry.”

Beyond handling routine tasks, AI may also be able to tackle more complex challenges such as forecasting, strategic planning, and data-driven decision-making. It can also supercharge creative endeavors like brainstorming sessions, where it can generate lists of fresh ideas quickly and effortlessly.

However, to milk AI’s potential, you need to use it regularly. Regular use allows leaders and their teams to understand AI’s nuances, discover its strengths and limitations, and build familiarity with its capabilities. This sustained engagement helps create a deeper intuition for where AI can add the most value, enabling leaders to integrate it more effectively into their decision-making processes.

“You won’t know what it’s good or bad at until you use it and you might find some surprising insights of what it can do well, and you might find some disappointments,” Mollick says. “That’s part of the process, so you have to use it to learn how it works.”

#2. Be the “human in the loop”

While AI can automate many tasks, it should never replace human oversight. For all their power, these systems still make mistakes. They are prone to presenting plausible-sounding but incorrect information, known as “hallucinations,” and tend to forget old information after learning something new, known as “catastrophic forgetting.” They are also bad at math (like, surprisingly so).

Imagine you ask an AI to create a financial analysis. It will produce a thorough-looking document. Even so, as CNET’s editors learned the hard way, that document may contain random mistakes and misinformation. Despite this, the model will present its analysis confidently and authoritatively, making those mistakes even harder to detect. AI plausibility leads to human gullibility.

“What is [AI] good at? What is it bad at? The problem is that if you don’t know that, you can fall prey to a phenomenon called ‘falling asleep at the wheel’ [where] you stop checking the work even if you know you should,” Mollick warns. “It’s easy to make mistakes if you don’t understand the shape of the [technological] frontier.”

For these reasons, Mollick’s second principle is to keep humans in the loop. Team members must scrutinize AI outputs, while leaders remain in charge to ensure the final decisions are based on sound information. AI can assist in gathering information, but human judgment ensures insights are accurate and aligned with organizational goals.

“What you need to do is sharpen your intuition by working with the tool,” Mollick says. “Get a sense of when you see something that might concern you.”

#3. Treat AI like a person (but remember it isn’t)

Mollick’s third principle encourages leaders to treat AI like a person by giving it context and clear instructions. If you’re seeking input for a marketing strategy, for example, prompt AI to adopt a persona — that is, to “act as a seasoned marketing strategist.” This persona helps tailor the AI’s responses to meet your specific needs.

But Mollick warns: “The danger of treating the AI like a person is you’re committing the cardinal sin of AI researchers, which is to pretend that a computer is a human. […] AI works a bit like a psychic. It’s good at figuring out what you want to hear and then telling you that. It’s easy to lose track of that if you start treating it like a human being.”

Remember that while AI may act like it understands a task, it isn’t reasoning and lacks comprehension of the outside world. It’s following statistical patterns found across billions of data points. Forgetting this fact can lead users to overestimate AI’s capabilities and be more persuaded by it. It can also blind you to its biases and limitations — such as generalizing based on data over-indexed by certain demographics or countries.

“Use it like a human, but always keep in the back of your mind that you’re talking to a machine. There’s no mind behind this. There’s no emotion or personality,” Mollick says.

#4. Assume this is the worst AI you’ll ever use

AI technology is evolving rapidly, and today’s systems, advanced as they are, will likely be surpassed in the coming years. As such, Mollick’s fourth principle is to assume this is the worst AI you’ll ever use. 

“We are in the middle of technological progress, not at the end. The progression of AI has been extraordinarily rapid, and it’s hard to measure directly,” Mollick says.

Leaders should see this as an opportunity to build foundational experience with the technology. They can then use that foundation to position their organizations to take full advantage of future advancements. As AI continues to improve, those who keep up will be better equipped to adapt and thrive.

Mollick adds: “We don’t know how far it’s going to go. We don’t know how good these systems are gonna get. We’re not in control of how fast these systems improve. We are in control of how we decide to use them and how we decide to apply them.”

Leadership in an AI-driven world

As AI’s influence on business practices grows, Mollick’s guiding principles provide a framework for embracing this shift. By adopting these principles, you not only learn how to utilize AI while maintaining the human touch that drives real innovation, you also retain a sense of control — at a time when many of us feel disempowered.

“As managers and leaders, you get to make these choices about how to deploy these systems to increase human flourishing,” Mollick says, adding, “AI is here to stay. You get to make the decision about how to handle, work with, and thrive alongside AI — rather than just be scared of it.”

So, take that leap, put in your lab hours of AI experimentation, and lead confidently with AI. You might be surprised by the insights and efficiencies it brings — all while staying firmly in the driver’s seat.