Sounding Right Is Not the Same as Being Right.
A well-articulated opinion and a correct one are not the same thing, and AI makes that gap easier to miss. On naming where a claim actually came from.
Dear younger me,
You’re going to have opinions. Strong ones. Well-articulated ones. People will nod when you speak. Slides will look convincing. Arguments will sound complete.
That’s not the same thing as being right.
Here’s the distinction you didn’t have words for yet: data is data, and opinions are opinions. Both matter. Confusing them is where credibility starts to leak.
Early on, this is hard to see because design environments reward clarity of expression. If you can explain something cleanly, it feels solid. If others agree, it feels validated. Meanwhile, the actual evidence is often partial, delayed, or inconvenient to surface.
So opinions start doing work they weren’t meant to do.
This matters because decisions don’t fail loudly. They fail later. Quietly. Downstream. Long after the meeting where everything sounded reasonable. When that happens, people don’t go back and ask how confident you sounded. They ask what you knew, and when you knew it.
AI makes this blurrier.
It can summarize, pattern-match, and articulate with confidence by default. It doesn’t hesitate. It doesn’t flag uncertainty unless you ask it to. The result sounds authoritative even when the underlying information is thin, inferred, or borrowed from somewhere else entirely.
That doesn’t make opinions less valuable. It makes honesty about certainty more important.
You’ll see how this plays out. Teams will justify decisions with language that feels data-driven but isn’t. Charts will be shown without context. Insights will be cited without origins. Over time, trust erodes not because people were wrong, but because they were unclear about what they actually knew.
If you don’t learn to label certainty accurately, a few things happen. Stakeholders stop distinguishing signal from polish. Feedback gets harder to parse. Your influence shrinks quietly, even if your output improves.
So slow down at the moment it matters most.
Before you speak, ask yourself where the claim came from. Did it come from observation, measurement, or repeated evidence over time? Or did it come from interpretation, intuition, or preference? Both are allowed. Only one is data.
Naming the difference doesn’t weaken your argument. It makes it trustworthy.
From the future,
Me