Video
Can technology change behavior? Can decision intelligence change the way we experience the world?
Apr 27, 2022

Some of the common misconceptions about AI and machine learning are that AI is going to actually have this real level of intelligence, but it's not. A lot of my friends are like, "Ooh, are machines going to take over the earth?" Not in my lifetime, not in my lifetime. We're so far from that. But that also gives these unrealistic expectations, even on professionals, when they're coming in and trying to understand what they can do with machine learning.

We are a very well-educated industry. The physicians and the people running the trials and the people treating the patients, the physicians, they really believe that they've got the right answers and the right information to treat their patients. When we bring, and we introduce some sort of machine learning into the process, it's difficult for them to really trust that. People tend to be skeptical of it. I don't think it's a fear. It's more a skepticism, but we do see time and time again, that it can pick the better sites. It can recommend the better drug that can tailor this thing for this patient to create better outcomes.

Everyone is out there working on machine learning and AI and figuring out how they can use AI, but they're not spending enough time on the human side of it. And the recipients and the users of these algorithms and these recommendations are people, and you need to understand that people make decisions in an often times not data informed way. They make decisions like humans do. And you need to acknowledge that. You need to lean into that and understand that as you're trying to display and put the information in front of them to optimize the process or the decision.

When you just start to display data, or you just start to make recommendations, you'll see that people just don't listen to them and you need to create some sort of user experience and you need to really start to lean into the psychology and sort of organizational processes of how the decisions are made. How do I, as a data scientist come in and say, "Oh, hey, start listening to this machine." What can I do drive that psychology of somebody that really knows the space and say, "I can actually make you better." That's the concept of decision intelligence… his concept that blends the decision-making, human psychology, the managerial sciences, and machine learning and user experience, to try to make sure that people actually make better decisions.

If they're going to decide I'm going to go forward with this molecule, or they decide we're going to spend more money on this advertising campaign or this marketing campaign, you need to understand who the decision makers are, why they make the decisions they make, what information they have to make the decisions they make, and then, is there space for machine learning to help augment that decision-making process?

A good example is, is there a default that we could provide to these decision makers saying, "Rather than providing you a bunch of information, you have to make some decision from scratch, can I say, well, here's the default. You're going to go to these countries for your trial. You're going to go to these sites, and this is, the machine picked these and provides the information as to why they've picked them" and then the people come in and start to decide, do I want to override some of these decisions? And those are some of the psychological ways that you can really start to understand and get past some of the adoption challenges, which really can then benefit both the decision-maker and the overall organization that's trying to make the better decisions.

My recommendation for how a company can best take advantage of decision intelligence is to start to acknowledge that accurate data science isn't enough. I think there's a lot of companies that are a little bit behind in thinking just having good algorithms and good data science is going to be enough and starting to measure, are people actually listening? Once people start to measure, are people listening, they'll realize that a lot of times they aren't and. When you can really start to study that and making sure that you're focusing on that and then augmenting it with the information needed to change people's decision-making processes, then you start to get the adoption of AI and machine learning that you really need to get the value from it.

Life sciences and healthcare, the decisions that we make are very critical. Obviously other industry's decisions are critical, but oftentimes not quite as life and death. We've spent a lot of time making sure that the experts in this field are very well-educated and so that means there's a very clear acknowledgement that these decisions are very important and need to be well-informed. That's why when we start to see decision intelligence and machine learning and the value that it can bring, if we can improve those experts' decision-making process, it becomes so critical and so valuable because you're literally saving lives.

Related solutions

Contact Us