When everything goes smoothly, we may be hollowing out the foundations of our decision making
It started with a dinner conversation extolling the virtues of AI in scheduling. AI can make everything seamless, saving time in the tedious back-and-forth among colleagues trying to set up a simple coffee date. But then a friend piped up: is there something we actually gain in that exchange, frustrating as it may be? Doesn’t it provide insights into how a person thinks and communicates? The conversation took a turn. In our pursuit of a frictionless society, made possible by AI, what are we giving up?
This is not a question that asks us to give up technology for routine tasks. Rather, it’s an invitation to think through the costs to individual growth and collective innovation, so that we can wisely design friction into an increasingly frictionless world.
My colleague Brinnae Bent, executive in residence in AI at Duke University’s Pratt School of Engineering, designed a DisagreeBot in her Trust Lab. When someone inputs “2 + 2 = 4,” the bot argues – using alternative mathematical frameworks such as non-Euclidean mathematics – that this might not hold true.
She wanted to highlight a fundamental limitation of current large language models: unless explicitly prompted otherwise, generative AI is designed to validate you. It tells you that your ideas are wonderful and your logic is sound. Life feels more pleasant when an intelligent system agrees with you. But constant agreement is not something we experience when interacting with fellow human beings – and for good reason.
AI is built on past probability, yet human progress is driven by imaginative improbabilities. It rarely follows the path of least resistance. We didn’t send astronauts to the Moon on “Yes” alone. There had to be friction. Friction highlights blind spots and missing perspectives, builds conviction by testing the strength of our arguments, defines a team’s values, and cultivates trust. It is only in moments of disagreement that we find out what truly matters to colleagues, bosses and stakeholders.
In the 1930s, Alfred P Sloan listened as his fellow General Motors board members quickly reached a unanimous decision on an important issue. Instead of calling for a vote, he stopped the room: “I propose we postpone further discussion to give ourselves time to develop disagreement and perhaps gain some understanding of what the decision is all about.”
Reconvening for the sake of disagreement might be a frustrating prospect. When someone stalls a meeting with a question that might alter the trajectory of your project when it’s close to the finish line, the natural instinct may be to move on; to just get it done. But Sloan had a point. Friction is the crucible of conviction. When you let something pass without careful debate, do you truly have conviction in the outcome? When we bypass the messiness of differing human perspectives, we save time but hollow out the foundation of our decisions. Conviction is forged in the moment when someone dares to say, “Wait” – forcing the team to defend and refine. Without that heat, the resulting product is brittle, destined to crack under real-world pressure.
So we must resist the urge to automate away the ‘messy’ parts of teamwork. Great teamwork comes from a willingness to sit with that messiness – because human beings are fundamentally messy. Friction sharpens vision. The value of working through disagreements to find collective clarity cannot be traded away.
Sanyin Siang is a Pratt School of Engineering professor and leads the Fuqua/Coach K Center on Leadership & Ethics at Duke University
