article
You Don't Have an Accountability Problem
By Abi Claus · LinkedIn · April 2026
Unlimited PTO, data filtered through fear, and why your AI investment might be the most expensive cultural mistake you've ever made.
The Problem
One of the sharpest young minds I know, a former colleague, texted me a scenario he was navigating, one that turned out to be a lot more common than he realized.
Engineers defaulting to the group on every decision. Nobody really owning an outcome, so nothing moves. Questions answered by committee. People afraid to share their own viewpoint because going against the group gets you ostracized. The person who dissents gets treated as a threat to the group's safety, so nobody dissents. And because nobody dissents, nobody actually owns anything either.
He called it an accountability problem. I stopped him right there.
What he described isn't an accountability problem. It's a fear problem. Every item on that list is a symptom.
Decision by committee so nobody gets blamed individually. Viewpoints suppressed because dissent is punished. Ostracism as the price of independent thought.
That's not a team that lacks accountability. That's a team that learned, through repeated experience, that standing out is dangerous.
You cannot fix a fear problem with an accountability solution. But that's exactly what most leaders try to do.
The Lie
Assign an owner. Problem solved. That's the accountability model most leaders are running.
It works just enough to feel right. Which is the dangerous part.
Think about unlimited PTO. A lot of companies offer it - sounds incredible on paper. Unlimited. No cap. Take what you need.
Except when you actually try to take a week off, there's no backfill. Your phone still buzzes. You take your laptop on vacation. You come back to a pile that makes you regret even leaving.
The policy says unlimited; the culture says something else entirely. Everyone knows the difference. Nobody says it out loud, because saying it out loud isn't safe either.
Put a name next to an outcome. Create a RACI chart. Someone's responsible now. That owner exists on paper but operates inside a culture that makes real ownership impossible.
Because real ownership requires something the org chart can't provide. It requires the person holding it to be able to tell the full story. The gaps. The unknowns. The parts that aren't working. The "I don't know" moments. The bad news before it becomes a crisis.
If that person has to filter the truth before it travels upward, if they have to manage how the news lands before they can deliver it, if saying "this isn't adding up" carries personal risk, then they don't own the outcome.
They own the exposure. Ownership without empowerment is performative accountability.
Your people aren't withholding information because they're bad at their jobs. They're withholding it because you built an environment where withholding it is safer than sharing it. And you, the leader receiving that managed story, are now making decisions on data that's been filtered through everyone's fear before it reached you.
You think you have an accountability problem. You have a filtered reality problem. And no RACI chart fixes that.
That's on the leader. Every time.
Why AI Makes This Urgent Right Now
For a long time this problem was manageable. Organizations moved slowly enough that the gap between what leaders knew and what was actually true could be absorbed. Bad data produced bad decisions, but the consequences arrived gradually enough to course-correct.
AI ended that grace period.
A senior technical architect I deeply respect says that every AI problem is really an integration problem. Yes, and. I'd add a layer underneath that. Every integration problem is really a culture problem.
AI is only as good as the data it runs on. And data is only as good as the honesty that produced it.
Yes, bad data has other causes too. Lazy inputs. Duplicate records. Systems nobody uses right. But fear-shaped data is uniquely dangerous because it's intentional, and it travels upward looking exactly like the truth.
If your people are filtering reality before it surfaces, if the culture rewards comfortable information over accurate information, AI doesn't fix that. It automates it. It runs your fear-shaped, perception-managed data through a system built for speed and scale, and produces fear-shaped outputs faster and with more apparent confidence than any human team could.
You haven't fixed the dysfunction. You've industrialized it.
Culture problems don't get solved by moving faster. They get exposed by it.
The leaders who haven't done this work are about to be very confused about why their AI investment isn't delivering what the pitch deck, or the POC, promised.
It always was a culture problem. It always will be.
