A report landed on the desk of an executive I know. It was a synthesis of ten thousand employee comments. Five themes. Clean bullets. A logical progression from grievance to solution. The AI had processed the mess of human dissatisfaction and returned a mirror that showed the executive exactly what they wanted to see: a solvable problem.

The problem is that the organization was actually on fire.

The sixth theme was missing. It was the theme of quiet, calculated exits by the top three percent of the engineering talent. It was the theme of a middle management layer that had stopped believing in the mission but started perfecting the art of compliance. The AI missed it because those people weren’t shouting. They were whispering in the gaps.

In its July 8, 2026 cover story, The Atlantic argues that we are entering a postliterate era. Most people can still decode words. Fewer can wrestle with them. We are outsourcing the intellectual struggle of synthesis to machines. We ask for the summary because we no longer have the stomach for the source.

The article is right about the cost of this convenience. When you stop wrestling with language, you stop thinking. But it misses the structural shift in power. The loss of the reading habit is a symptom. The real shift is a new, invisible divide in leadership.

On one side stand the consumers of compressed reality. On the other stand the few who still know how to read between the lines.

The Fiction of False Coherence

AI is a master of the smooth surface. It takes the jagged, contradictory, and often nonsensical reality of a modern corporation and irons it flat. It gives you a world that makes sense.

Human organizations are not logical. They are collections of egos, histories, terrors, and unspoken pacts. When an AI summarizes a department’s failure, it might point to “misaligned incentives” or “process friction.” It will rarely tell you that the VP is terrified of the CEO, and therefore everyone below the VP is lying to stay safe.

The AI makes everything sound like a McKinsey slide. If you lead based on that summary, you are making decisions on fiction. You are managing a ghost ship while the real vessel is taking on water.

This is the trap of the dashboard. Senior leaders are increasingly removed from the raw data of their own companies. They receive summaries of summaries. They see five themes because the system was designed to find five.

Who catches the sixth?

The New Cognitive Inequality

We are heading toward a world where discernment is the ultimate currency.

It is a mistake to think AI will democratize leadership. It may do the opposite. It will provide a cheap, high-speed way for the uncritical to be wrong more confidently. The gap will widen between those who can detect what was flattened and those who take the summary as gospel.

Junior employees are already being fed summaries as a matter of course. They are losing the “apprenticeship of the raw”: the years spent reading the full customer transcript, the entire legal brief, the unedited field report. Without that exposure to the unpolished truth, they cannot develop the scar tissue that we call judgment.

The leaders who thrive will be those who intentionally break the compression. They will be the ones who demand the “weird” data. They will look for the outlier comment that doesn’t fit the five themes.

In my book, The Bias Advantage: How Unconventional Leaders Gain Power in an AI-Driven World, I talk about the power of pattern recognition across contexts. Unconventional leaders: those who have often existed on the margins of corporate power: have a unique edge here. If you have spent your life reading between the lines because the official story never quite included you, you are already trained for this moment. You know how to spot the thing that was omitted. You know how to hear the silence.

How to Read Against the AI

To lead now, you must learn to read against the machine. The point isn’t to reject AI. The point is to exercise a higher form of literacy — reading against the machine.

First, stop looking for consensus. If your AI summary says “the team is generally optimistic,” ask for the three most cynical comments. Consensus is where nuance goes to die. The most valuable signal in your company is often the one that the majority is trying to ignore.

Second, watch for the “missing 6th theme.” If the output feels too clean, it is likely incomplete. Real human problems are rarely divisible by five.

Third, protect your own discernment. You cannot outsource your thinking and expect to keep your edge. Spend time with the raw. Read the full letter. Listen to the unedited recording.

The advantage belongs to the person who can see what the machine has smoothed over.

At LBZ Advisory, we help leadership teams move beyond the performance of AI and into the reality of it. We help boards and CEOs identify the high-leverage bets that actually scale, rather than just generating more noise. This requires more than a new tool. It requires a transformation in how you see.

The age of easy reading is over. The age of deep discernment has begun.

The employee willing to tell you that the AI is wrong: and why: is the most valuable asset you have. If you have built a culture that rewards the summary over the truth, you will never hear them.

You can find more on these strategies in The Bias Advantage, available now on Amazon. For teams looking to scale these insights, we offer bulk purchase options and executive briefings through our VIP list.

The machine can summarize the world. It cannot decide which part of it matters. That remains your job.