Issue 150  July 31st 2026

AI raises the floor. It may also lower the ceiling.

The usual argument for AI-assisted creativity is straightforward: it helps people generate more ideas, overcome a blank page, and improve weaker first drafts.

That argument is largely true, but also incomplete.

Researchers writing in MIT Sloan Management Review synthesized four studies involving short-story writing, circular-economy solutions, humor, and collaborative storytelling. Across those settings, AI assistance improved individual creativity and productivity, but reduced the diversity of ideas across the group. Average quality rose while the number of unusual outliers fell .

AI raised the floor, but it also lowered the ceiling.

The mechanism was not that AI lacked ideas. It was that people anchored on its suggestions too early. A plausible option arrived before they had explored enough of the search space themselves. Once something looked good enough, the work shifted from discovering possibilities to improving the first acceptable one .

That pattern starts before creativity.

An article on the curiosity cycle describes five connected stages: recognizing an information gap, building anticipation, exploring, forming longer-term memory, and experiencing the reward of discovery. AI can compress that entire sequence into one quick exchange, especially the opening period when you notice uncertainty and decide whether to explore it.

The answer arrives and the curiosity barely gets started.

Critical Judgment faces the same pressure. A recent article in The Conversation describes critical thinking as two linked moves: reflection, followed by judgment. Reflection is the pause in which you question evidence, examine assumptions, compare interpretations, and recognize the limits of your perspective. Judgment turns that reflection into a decision or action.

AI makes it easy to jump directly to the second move.

A polished response looks ready to judge. But without the reflection beforehand, you may only be choosing among the frames the tool handed you.

That is the common thread:

• Curiosity needs time for the information gap to remain open.
• Creativity needs time for possibilities to diverge.
• Critical Judgment needs a pause before forming a conclusion.
• AI makes each of those intervals optional.

And the market may already be distinguishing between people who preserve that judgment and people who do not.

Upwork’s 2026 data found that freelancers doing generic generative and creative AI work saw average earnings per contract decline, while complex AI-augmented professional services combining technology with domain expertise and judgment grew in both volume and earnings.

AI fluency alone is not the premium. What you contribute beyond the output is.

THE TELL

Here is how to recognize when AI is improving the immediate work while narrowing the capability underneath it:

• Your first response to uncertainty is to prompt, not explore.
• Your team produces more polished ideas, but they increasingly resemble one another.
• The first plausible suggestion becomes the direction before anyone has framed alternatives.
• You evaluate the answer but never question why you received that particular answer.
• You can revisit the AI output, but you cannot reconstruct the reasoning without it.
• The work moves faster while your own position becomes harder to explain.

The danger is not that AI gives you a bad idea. It is that AI gives you a good-enough idea early enough to prevent a more original one from emerging.

YOUR PRACTICE THIS FORTNIGHT

🛠 The Human-First Rep: Diverge Before AI

Choose one problem you need to solve this week: a recommendation, concept, strategy, workshop, product decision, or difficult piece of writing. Before opening an AI tool, spend ten minutes generating at least five different possibilities. Do not polish them. Do not rank them yet. Aim for range.

Then ask:

• What assumption does each option make?
• Which options are merely variations of the same idea?
• Which one feels least obvious?
• Which possibility would disappear first if AI offered a plausible answer now?

Only then bring in AI.

The goal is not to prove that your ideas are better. It is to give your curiosity and judgment enough time to establish a wider search space before the tool begins influencing it.

🤖 The Sparring Prompt: Expand the Range

Paste your human-first options into your AI tool:

I generated these possibilities:

[paste your options]

Do not choose the best one yet.

First, identify where my ideas cluster around the same assumptions, definitions, or constraints.

Then propose three alternatives that use genuinely different framings of the problem. Explain what makes each alternative structurally different, rather than merely another variation.

Finally, identify one unusual option from my original list that may deserve more attention before I converge on a direction.

Compare the new options with your original set.

Did AI widen the search space, or pull the ideas toward a polished middle?

That is PAIR in practice: Prime with your own possibilities. Assess the range with AI. Interpret what is genuinely different. Retain responsibility for the direction.

Think first, then consult.

THE ARTICLES

📄 How AI Is Shortening Your Curiosity Cycle | Forbes

The useful idea here is that curiosity is a cycle, not a personality trait. Recognizing an information gap, exploring it, and experiencing discovery all exercise capabilities that an instant answer can bypass. The recommendation is not to avoid AI. It is to form possibilities, predictions, or better questions before using it, and to continue asking what remains unknown after it responds.

📄 The Hidden Cost of AI-Assisted Creativity |MIT Sloan Management Review

Across four studies, AI assistance improved individual creative performance while narrowing the diversity of ideas produced collectively. The important variable was sequence: human-first ideation preserved more diversity than bringing AI into idea generation at the start. For leaders, that means measuring more than average quality. Look at the range of ideas, the assumptions represented, and whether unusual outliers survive long enough to be considered.

📄 New Data: AI Is Making Your Human Skills More Valuable | Forbes

Upwork’s data shows a widening gap between generic AI production and higher-judgment work. AI-related freelancers earned an overall hourly premium, but undifferentiated generative and creative work saw lower per-contract earnings, while AI-augmented professional services combining domain expertise with AI grew substantially.

The market signal is not simply “learn AI.” It is: use AI inside expertise that helps you frame the problem, judge quality, and connect execution to an outcome.

📄 Critical Thinking Has Become an AI-Era Buzzword | The Conversation

The piece offers a useful correction to vague calls for more critical thinking. Reflection comes before judgment: pause, examine the assumptions and evidence, consider why this particular information reached you, and only then decide what deserves confidence or action. The AI-era upgrade is not only questioning the answer. It is questioning the framing that produced the answer.

ONE THING TO TRY

Before your next important AI interaction, create three possibilities of your own. Not because three is enough. Because zero leaves the tool responsible for defining the range.

Then ask AI to challenge and expand what you started - not replace the exploration that should have come first.

Reply and tell me: Where would a “diverge before AI” step make the biggest difference in your work? I read every response.

The human edge is not automatic. It has to be built.