This 'Clairvoyant' Thought Exercise Can Help You Make the Right Choice in Just Seconds, Says Google's Former Chief Decision Scientist Cassie Kozyrkov, leader in decision intelligence and CEO of Data Scientific, delves into effective decision-making and its role amid the rapid advancement of AI.

We've all been there. It's time to make a decision, and you're stumped. Which choice is the right one? Whether it's a big decision or small, feeling a sense of analysis paralysis isn't uncommon, even among so-called "experts" in making tough calls: 57% of newly appointed executives said that decisions were more complicated and difficult than they expected, research from Harvard Business School found.

Just like anything else, decision-making is a skill that can be learned and honed for better results — inside and outside of the workplace. But too often, the decision-making process isn't given due attention; instead, it's left up to the highest in a chain of command who doesn't fully understand the implications of a choice, sometimes to the detriment of the whole.

"People treat decision-making as something that just happens, and there's also such a hierarchical aspect to it," Cassie Kozyrkov, former chief decision scientist at Google and leader in the decision intelligence space, explains. "When it's so tied in with hierarchy as opposed to skill, you do get a little bit of overlooking the fact that there are skills that you could bring to it."

Related: How to Prioritize When Making Decisions as an Entrepreneur

^{Image Credit: Courtesy of Persona PR}

"Decision intelligence is [about] turning information into better action at any scale, in any setting."

Kozyrkov, who holds graduate degrees in mathematical statistics, psychology and cognitive neuroscience from Duke University and North Carolina State University, helped train and build data science teams at Google. These days, she's a trusted advisor to business leaders at Fortune 100 companies; an outspoken advocate for the need to "stand up to AI," an issue of negligent human decision-making at scale; and the CEO of Data Scientific, which provides expert advisory and consulting services for AI.

"For me, decision intelligence is [about] turning information into better action at any scale, in any setting," Kozyrkov says. "There's the stuff that happens to you that you, by definition, can't do anything about; [you] can't see it coming. And then there's what you do have control over — and that's the quality of your decisions."

Related: Let Go of These 10 Things and Start Making Better, Faster Decisions

In her Decision Intelligence LinkedIn course, Kozyrkov breaks down the science of decision-making. The first part delves into the basics of improving technique — no special tools or resources required. One of them relies on a decision science concept known as the value of clairvoyance, Kozyrkov says. "This isn't about crystal balls or psychics," she adds. "It's a thought exercise people forget to make, particularly in the data setting."

The exercise goes like this: Imagine that you have access to an all-knowing psychic who could give you all the information you need to make the best decision, then consider how much you'd pay them for those details. "If you find yourself answering with a small number, that is a strong signal to you that you shouldn't be investing in a whole lot of data, going through a whole lot of process," Kozyrkov says. That's when a "go with your gut" approach can be more effective than a deep data dive.

"If you know these three things, you know a lot about the [AI] system."

With the rapid development and adoption of AI, Kozyrkov is particularly interested in "the decision-maker behind the curtain." When the technology isn't used for basic organization or record keeping but instead to automate decision-making at scale, leaders have the potential to "impose [their] will on a million decisions," Kozyrkov says. That's when it's crucial to analyze precisely how data is used and decisions are made.

Kozyrkov has developed a set of questions to help people gain that understanding; she calls them the "Kozyr criteria." "[They're] the decisions that create an AI system," she explains. "If you know these three things, you know a lot about the system. You know a lot about its risks and how it might benefit some people more than others, for example. And you know a lot about what could go wrong with it right off the bat."

The first question to ask is, What is the objective of the system? "That is very much a subjective question that a decision-maker has to answer," Kozyrkov says. "There's no one true, right answer for it."

Next up, consider the provenance and schema of the data. "Of course, we would love to go and examine your data set," Kozyrkov says, "but it's not like you're going to show it to us. So tell us please where you got it, how you got it, perhaps how you checked it. Did you have a diverse set of individuals checking it for problems? What do you know about where it came from? How much are you relying on it?"

Finally, How do we know that it works? "What were the criteria, and under what conditions was it tested? If you are told that some food is safe and you find out that it's only ever been eaten by cats, you might worry about whether it's safe for you as a human, but if no one tells you how it was tested and where and under what conditions, you have very little to go on," Kozyrkov explains.

Related: These Decision-Making Tactics Can Help You Formalize Your Process and Make Better Choices

"It always comes from humans. There's no way to remove the human."

Kozyrkov admits that these criteria are "deeply subjective" and dependent on what the decision-makers in question are trying to achieve. However, allowing AI to be "faceless" and ignoring the fact that people are sitting in a room and influencing the technology's behavior opens up the potential to harm billions of people.

"It always comes from humans," Kozyrkov says. "There's no way to remove the human, but sometimes you just scale up a human so much that it boggles the mind that people can be scaled up that much — [and] we forget that there are humans behind it. [That's why] I would love to have a decision-intelligence-first approach."