Bloomberg reports on a split the Wharton-GBK Year Four study surfaces between employees and their managers: if people are being encouraged to experiment with AI, who takes the blame when something goes wrong?
Mindspan Labs founder Jeremy Korst, a co-author of the study, on where most organizations actually are:
"Organizations are learning in real time. When I ask this question of a senior executive, the most frequent answer I get is that they hadn't thought about it yet."
The way we see it, that answer is the finding. Accountability for AI outcomes is a Ring 2 question — decision rights and who owns the result — and it is the ring that turns ambition into shipped work. An organization can encourage experimentation and still have no answer to what happens when an experiment is wrong, and the gap does not show up as a failure. It shows up as hesitation: people decline to use the tool on anything that matters, because the consequences of being wrong are undefined and therefore theirs.
So what: Decide who owns an AI-assisted decision before you encourage more of them. Not a policy document — a named owner, a review path and a stated tolerance for error. The absence of an answer is itself an answer, and your teams have already worked out what it is.