AI’s Kobayashi Maru
Everyone’s a Manager Now, and Most Aren’t Managing
Image: Depositphotos
This Member Article is authored by think tank member Christine Heckart, CEO, co-founder of Xapa, which uses game mechanics to help people build the human skills for success in an AI world, in 5 minutes of play a day. This is especially important in an AI workplace where everyone is a 'manager now' of AI teammates.
In 1969, the psychologist B.F. Skinner wrote: "The real question is not whether machines think but whether men do."[i]
Artificial intelligence was still in its infancy when Skinner raised a question that remains remarkably relevant today. He was worried that as machines got better at producing output, people would quietly stop doing the harder work of thinking, and nobody would notice the swap until it was done.
Fifty-odd years later, we've built the machines. Now we get to find out whether he was right.
The $200 billion shrug
Some companies may spend something like $200 billion on AI this year. McKinsey's research finds that only about 6% of surveyed organizations qualify as AI high performers, achieving significant value and measurable bottom-line impact. Look closely at the winners, and McKinsey's research emphasizes organizational transformation, workflow redesign, and the human capabilities needed to capture AI's value.[ii]
Translation: four out of five reasons an AI project works has nothing to do with which model you bought. They come down to whether the person at the keyboard knows what to ask, what context to give, how to judge the answer, and when to ignore it entirely.
Yet we keep measuring AI progress by logins and seat licenses. That's like judging a restaurant by how often the stove is on. This was never a technology problem. It's a thinking problem.
Everyone is a manager now
Your interns have interns. Your frontline employees have frontline employees. Even your executives are directly "managing" AI agents – which is like going back to directly managing interns again.
Each AI agent is like a new hire with two PhDs and no context, wisdom, or business judgment. It has read more than any human could in a hundred lifetimes: truths, lies, and half-truths, with no reliable way to tell them apart. And boy, is it productive.
AI will produce a memo, a strategy, or a customer email in seconds, beautifully formatted and supremely confident. It is often wrong. Wrong in part, or in whole? Only you can figure that out. It doesn't know your biggest customer is quietly furious, or that the clever fix it just proposed blew up in your face two years ago. Push back, and it cheerfully agrees and tries again. And again. Until you make a call.
So, we need to teach critical thinking and judgment, broadly and fast. Here's the catch: everyone is too busy to stop and learn. Skip it, and we keep scaling nonsense. It's a Kobayashi Maru, the Star Trek training test with no winning move.
Captain Kirk beat it by changing the rules. So can we.
The old rules don't work anyway. One-time classes and hour-long videos don't stick. People forget most of what they hear within days, and behavior barely moves. Habits change the way they form: short doses, delivered continually, reinforced by managers in both what they say and what they do.
Even five minutes a day of focused, reinforced practice can be a meaningful start. That isn't much. You know what's less? Zero minutes. Short bursts of practice that build the judgment, thinking, and management skills everyone needs now. Without it, we drown in mediocrity: an ocean of beautifully formatted slop, written by machines, for machines, with no human left to notice.
The skill under all the skills
Directing AI is only half the job. The other half is what AI can't do yet: lead, build trust, persuade, negotiate, and own the outcome. Our customers are still human, and dealing with humans still takes one.
Underneath all of it sits one skill we barely notice because it's so ambient: reading context. AI can record every word in a meeting and still miss what happened in the room. It can't know the tensions, history, and politics shaping your decision unless someone tells it. For now, that someone is us.
The most valuable intelligence in any company isn't artificial. It's the authentic kind: ours. Skinner's question still stands. The machines are thinking. Are we?
NOTES
[i] B. F. Skinner, Contingencies of Reinforcement: A Theoretical Analysis (New York: Appleton-Century-Crofts, 1969)
[ii] Dan Tinkoff et al., “The State of AI: On the Road to ROI,” QuantumBlack, McKinsey & Company, August 25, 2026, https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai.
Christine Heckart
Author
Christine Heckart is CEO and co-founder of Xapa, which uses game mechanics to help people build the human skills for success in an AI world, in 5 minutes of play a day.
Christine has experience on both sides of the board table. She brings over twenty five years of experience in the technology industry having held a variety of leadership roles including CEO, GM, president, and CMO in industries that include entertainment, SaaS, data analytics, security, networking, and storage. She has worked in B2C and B2B business models at companies ranging in size from one person to over one hundred thousand at iconic brands such as Microsoft, Cisco Systems and NetApp. Having begun in new product development, Heckart's passion is in forging wholly new markets, and she has launched three new categories that have reached several billion in revenues annually. She has also created disruptive business models and worked through big market downturns and inflections.