An old story about Jeff Bezos that I read years ago resurfaced in my memory when I learned that Amazon just abruptly changed course and decided to shutter its restaurant delivery business. According to the teller of this tale, Bezos once visited Basecamp to give a talk and take questions from employees there. At one point, Bezos described the kind of people who are “right a lot.” The smartest people, he said, are always revising their thinking. They might think one thing today that they’ll readily contradict tomorrow because they have new data to support a new opinion.


Other luminaries have shared this opinion. Science fiction author Vernor Vinge, in his book Fire Upon the Deep, wrote “Intelligence is the handmaiden of flexibility and change.” Salesforce CEO Marc Benioff was quoted as saying, “You must always be able to predict what's next and then have the flexibility to evolve.”


I wholeheartedly agree with this line of reasoning, and I see direct connections to machine learning.


We form opinions based on data and experiences that we’ve accumulated to-date. New events occur, and we gather new data and ideally revise our understanding, so we can seek new, better solutions to old problems. This is exactly what machine learning does. Business conditions change, and the shift impacts our objectives: What was once considered a lucrative investment in a delivery service could be considered a poor investment tomorrow. A low-priority target can suddenly become a high-value target as the market shifts or a new drug launches. Quick responsiveness to change is an ideal human characteristic, but most of us have a hard time with it; machines do not. AI can respond to events at the pace of change.


Haven’t we seen bias in AI and machine learning? Yes, but in those cases, we’re seeing human bias embedded in algorithms, and it’s a challenge that is being overcome. In approaches that leverage ensemble machine learning models, multiple models are given an opportunity to weigh in with different answers. Many different perspectives are considered, and they contribute relative to their perspective on the problem being solved. It’s like having a team of experts with diverse backgrounds considering one question and coming to the most sensible answer together.


I believe we will always rely on human intuition and creativity in some measure, but we’re entering an age in which AI and machine learning will have a seat at the table, even during some of our most complex decisions. We can look to this change with trepidation, or we can embrace how it will enrich our lives. Each of us will soon be able to lean on a team of trusted, unbiased and intellectually flexible advisors—personalized solutions that understand our role and leverage our company’s data—that will always be ready to support our decisions. And these advisors’ perspectives will evolve at the inhuman pace of change that is to come.