The Work Gets Easier. Does the Thinking Get Harder?

The Work Gets Easier. Does the Thinking Get Harder?

There is a particular satisfaction in making something difficult look easy. The presentation finally comes together. The code works. The analysis is complete. The proposal is beautifully written. The campaign launches. The prototype functions.

Behind the finished thing may be days, weeks or months of research, production, revision, coordination and problem-solving.

Generative AI is collapsing some of that distance. This is usually described as a productivity story. And it is one.

• But when producing the work becomes easier, something else becomes harder to hide: the quality of the thinking behind it.

When execution gets easier, intention becomes more visible

Much of organizational life revolves around producing things: decks, reports, emails, briefing notes, campaign assets, research summaries, spreadsheets, code, plans and strategies.

AI dramatically reduces parts of that production cost. Which sounds liberating. But remove enough friction and a slightly uncomfortable question appears: What were we actually trying to do?

The question is no longer merely Can we make the deck? It becomes What decision is the deck supposed to enable? Not Can we build the feature? but Should this feature exist?

Execution used to consume enough effort that completing the artifact itself could feel like progress. AI makes it easier to discover whether it actually was.

A sophisticated answer can still begin with the wrong question

AI is very good at helping us answer the question we give it. That means the quality of the output depends partly on something upstream: Who framed the problem?

• A poorly framed question can produce an extraordinarily sophisticated answer. A false assumption can become the foundation for a beautifully reasoned recommendation. And because the output can look so polished, we may become more susceptible to mistaking coherence for correctness.

This is why prompting, while useful, feels like only the surface of the challenge. The deeper capability isn’t merely knowing how to ask AI for what you want. It is knowing whether what you want to ask for makes sense.

More options make selection more important

AI can widen the field of possibility enormously. That is valuable. It can challenge premature conclusions and surface alternatives we might not have considered.

But at some point, the constraint is no longer ideation. It is selection. Which option fits the situation? Which trade-off are we willing to accept? Which possibility contradicts something important we know from experience? Which idea should we abandon even though we personally like it?

• Generating possibilities and exercising judgment over possibilities are different capabilities.

When a decision matters, an organization still needs to know where judgment resides—and what qualifies someone to exercise it.

But making and thinking were never separate

There is a risk in talking about AI and knowledge work as though tasks can be neatly separated into “thinking” and “doing.”

They can’t.

A designer moves elements around and suddenly sees that the hierarchy is wrong. A writer begins a paragraph and discovers that the argument doesn’t hold. A developer encounters an implementation problem that reveals a flaw in the original concept. A strategist maps a system and notices a relationship nobody had articulated.

• The hand is not simply executing instructions received from the brain. Production itself can be a form of inquiry.

Making produces information. You're not merely translating a fully formed idea from your head into an artifact. The artifact talks back.

You put two things beside each other and discover a relationship you hadn't anticipated. You try to execute an idea and discover it was structurally impossible. You sketch something and suddenly understand the problem differently. You write the sentence and realize you don't believe it.

In other words, execution isn't downstream from thinking. It is one of the places thinking happens.

• If we automate more of the making, we need to ask not only: What labour have we removed? but also: What forms of thinking used to happen inside that labour?

THAT is a much more consequential question. Because perhaps some friction really is useless and should disappear. But some friction is epistemically productive: we discover things because we have to wrestle with the material, the sentence, the code, the system, the prototype, the people.

And therefore the interesting design problem isn't simply how much execution can we automate? It's: Which parts of making are merely costly—and which parts are how we come to know?

• Some friction is epistemically productive: we discover things because we have to wrestle with the material, the sentence, the code, the system, the prototype, the people.

Production itself can be a form of inquiry.

Production itself can be a form of inquiry.

Production itself can be a form of inquiry.

This complicates the appealing idea that AI should simply automate execution so humans can move “up” into higher-order thinking.

Sometimes the thinking happened because we were doing the work. A practitioner develops judgment partly through repeated encounters with material, constraints, people, mistakes and consequences.

• Expertise accumulates not only from knowing what should happen, but from having watched what actually happens when an idea meets reality.

If AI removes us from parts of production, the question isn’t only what labour we save. It is what forms of learning may disappear with it.

Perhaps the challenge is preserving the feedback loops through which expertise develops.

Faster execution can make bad direction more expensive

Suppose a team misunderstands a problem but moves slowly. There is time for friction to intervene. Someone questions the brief. A technical constraint exposes a contradiction. A stakeholder reacts unexpectedly. A prototype fails.

Now imagine the same misunderstanding with dramatically accelerated execution. The team can generate the campaign, prototype the product, automate the workflow and deploy the system before anyone has seriously interrogated the premise.

Efficiency magnifies direction.

• AI doesn’t merely help us move faster. It increases the consequences of knowing where we’re going.

This may make an unfashionable organizational capability newly valuable: the ability to stop. To question. To reframe. To recognize when speed is amplifying the wrong decision.

“Human in the loop” doesn’t tell us whether the human has judgment

“Human judgment” is becoming one of those fashionable phrases that appears everywhere in discussions about AI. Keep a human in the loop. Use human judgment. Humans provide context.

All sensible. But putting a person into a workflow does not automatically produce judgment.

A human can approve something without understanding it. Rubber-stamp a recommendation because it arrived wrapped in persuasive language. Be biased, rushed, inexperienced or wrong.

• Judgment has to be developed. It draws upon knowledge, experience, context, pattern recognition, values and evidence.

So perhaps the important question is less: Where do we keep a human in the loop? and more: What must that human be capable of noticing?

Accountability changes the character of a decision

There is another reason some decisions cannot be reduced to prediction: Consequences belong to someone.

An AI system can produce a recommendation. It does not bear responsibility for what happens when that recommendation becomes real.

Organizations do. People do.

This doesn’t mean AI should be excluded from consequential decisions. Its ability to surface evidence, challenge assumptions and model scenarios may help people make better ones.

But analysis and accountability are not the same thing. There remains a difference between producing a recommendation and being responsible for what happens next.

Expertise may become harder to see just as we need more of it

A novice and an expert can use the same model. Both can receive polished answers. But the expert may notice something the novice doesn’t: a missing assumption, an impossible recommendation, a subtle contradiction, a context-specific exception, a plausible answer built on the wrong premise.

Expertise changes not only what someone can produce: It changes what they can see.

If we evaluate capability primarily by the quality of the visible artifact, AI may make expertise harder to recognize precisely when expertise is becoming more important elsewhere.

• We may need to become better at recognizing the less visible capabilities surrounding the artifact: problem-framing, interpretation, discernment, direction, critique and decision-making.

So does the thinking actually get harder?

Not necessarily.

AI can make thinking easier too. It can challenge an argument, expose alternatives, summarize unfamiliar fields, find contradictions, structure a messy problem and give someone a thinking partner they might never otherwise have had access to. Used well, it can extend human reasoning.

So perhaps “harder” isn’t quite the right word. Perhaps the burden placed on thinking increases. When execution becomes cheaper, organizations can attempt more things, produce more things and make more decisions. That increases the importance of choosing what deserves to be made.

• Producing something and knowing what is worth producing are not the same capability.

Neither are generating an answer and recognizing a good one. Neither are moving quickly and moving wisely.

AI may make the work easier. It may even help us think better. But once we can make almost anything—we have to get much better at deciding what is worth making.

::::::::::::::::::::::::::::::::::::::::

Contemplations & Provocations | A field essay by
Marie Fe Isla Rae

© 2026 Marie Fe del Rosario. All rights reserved.
Terms of Use · Privacy Policy · Permissions & Collaborations

All content on this site—including text, frameworks, methodologies, visuals, and strategic concepts—is protected by copyright and may not be reproduced, adapted, or used without prior written permission.

Authorship and web design by Marie Fe.

Field notes on consequential moments, creative strategy, experience, narrative, design—and the work of making possible futures tangible.

For people interested in what happens between imagining and making.