Hey, Nathalie! I this this is pretty easy to spot once you know what to look for. I usually find the answer in the small decisions, not the big announcements.
-Look at what gets rewarded IN PRACTICE. If the company says they value “thoughtful decision making” but promote the person who ships the fastest? They value speed. If the company says they value “collaboration” but the biggest bonuses go to the individual contributors who gatekeep information - they value solo performance. What gets rewarded is what’s actually valued.
-Watch what leaders do when there’s a conflict. If someone experiments with AI, it fails, and costs the company some time and a little money. Do the leaders get curious (“what did we learn?”) or punitive (“why did you waste time and resources?”); that moment tells you everything about whether they actually value experimentation or just say they do.
-Check the budget. Values are cheap until they cost something. Company says they value learning and development and employee growth but there’s no training budget and people can’t take company paid time off to learn? Not really a value. They say they value work-life balance but there’s no coverage plan so people can actually disconnect? Not serious about it.
-Look at who gets hired and who gets fired. If your values say “customer obsession” but you keep hiring people with no customer-facing experience and letting go of your best support folks, your real value is efficiency or cost control. Hiring and firing decisions reveal what you actually can’t live without.
Notice what’s okay to say out loud versus what’s only said in private. If people can openly say “i tried this and it didn’t work” in meetings, you value learning and experimentation. If that only happens in whispers, you value the appearance of success.
The gap between narrated values and real actioned values is where almost all of the dysfunction lives.
This is such an important question, and honestly, I don’t have a perfect answer. But I’ll share what I’m seeing.
You’re right that AI accelerates everything - including the bad stuff. If a company was already cutting corners or optimizing for short-term metrics over long-term trust, AI just makes that faster and less visible. You can generate misleading content at scale. You can automate decisions that harm people. You can create distance between leadership and impact.
Why this happens:
Leadership isn’t usually trying to be unethical. They’re just operating under enormous pressure - from boards, from investors, from competitors. And when you’re stressed and the goal is clear (hit the number, ship the thing, beat the competitor), it’s easy to lose sight of how you’re getting there.
Add AI into that mix and you get even more distance. The person making the decision doesn’t see the customer harmed by the automated rejection. The executive doesn’t see the content that feels manipulative because it was AI-generated at scale. Distance creates ethical drift.
What actually helps (in my experience):
Make impact visible. When leaders are far from the consequences of their decisions, bad things happen. Create rituals that force them to see it. Customer calls. Support ticket reviews. Whatever brings them closer to what’s actually happening.
Slow down the feedback loop. AI makes it easy to ship fast without seeing what breaks. Build in checkpoints. “We’re piloting this AI feature for two weeks, then we’re looking at what it actually did.” Don’t just automate and assume it’s fine.
Empower people to say no. If the only way to hit targets is to do something sketchy, and nobody can push back, you’re going to get sketchy behavior. Create real channels for dissent. Protect the people who raise concerns.
Hire for values alignment, fire for values violations. This sounds obvious but most companies don’t actually do it. If someone hits their numbers but does it in a way that erodes trust, and you promote them anyway - you just told everyone what you really value.
Small teams, clear ownership. The bigger and more diffuse the organization, the easier it is to make bad decisions because nobody feels personally responsible. Keep teams small enough that people feel the weight of their choices.
But here’s the hard truth: If the incentive structure rewards short-term extraction over long-term trust, no amount of values statements will fix it. The system will optimize for what it rewards.
So sometimes the answer is - you can’t fix it from the inside. You can create pockets of integrity in your team. You can be the leader who refuses to take the shortcut. But if the whole organization is optimizing for the wrong thing, you might be fighting a losing battle.
The question then becomes: how long do you stay and try to change it versus when do you leave and build something different?
Coming back to this thread @stephanie_muxfeld and @Kike_Pena , seems this thread is more relevant than ever, and in particular, I’m seeing the need for stronger systems that integrate AI thinking into the workflow. Without this, there’s an asymmetric force that starts to emerge. Curious where you see things headed.
