How to Build and Deliver AI Worth Trusting
All work in AI begins with a question. For many leaders, this might be “How do I get the most out of our AI?” or “How do we get more of our people to trust the AI”? Both questions tackle the concerns around adoption and efficiency, but often miss the essential work in controlling for risk. The better question is the one centred around the human experience: “How can I build and deliver AI worth trusting?”.
Let’s start by acknowledging that trust and trustworthiness are not the same thing. Trustworthiness describes the qualities that make a system deserving of your trust: whether it can perform its intended task, behave consistently, and serve your interests. Trust on the other hand, is a person’s willingness to rely on that system. Trustworthiness is what happens next: using the output, following the recommendation, delegating the decision.
Consider a bridge. Its load-bearing capacity is a fact about its design, materials and maintenance- the bridge’s qualities. A person may feel uneasy about crossing a bridge that is safe, or confident about crossing a bridge with structural weaknesses.
Either way, their reaction to the bridge does not change the engineering. In the same way, confidence in an AI system is not evidence of its reliability.
This distinction matters because organizations will often measure the wrong thing and then make decisions based on the wrong results. Positive user sentiment, high adoption and a smooth interface tells us about the qualities of experience. It will not tell us whether an AI system is accurate, robust and appropriately aligned to its purpose.
The better executive question is not simply, “Do people trust our AI?” It is, “What evidence shows that our AI deserves their trust?” That changes the conversation from persuasion to evaluation, and from perceived safety to demonstrable performance.
Building confidence may be part of responsible adoption, but confidence should follow evidence, not be a substitute for it.
Starting November 2, our five-part series will explore these distinctions and share practical guidance to help you build an AI bridge that is both trustworthy and sustainable. Join our community on LinkedIn to follow the series.
Adapted from: Kathleen McGinn and Michael Pascu, “Stop Building Trust in AI. Start Building AI Worth Trusting,” published June 9, 2026