Output vs Ownership: What AI Can’t Do for Your Team

I’m very pro AI and believe it’s a major unlock for humanity. I also got it wrong for a while.

I thought AI would take over a lot of the work people do in a business. I don’t think that anymore. AI can produce a lot of work fast. But the person using it doesn’t always understand that work or stand behind it.

That’s the real difference between AI and a hire. You hire a person to own the work, not just to produce it.

Ownership is knowing a piece of work well enough to explain it, defend it, change it, and take responsibility for its results.

Why Doesn’t Faster Output Mean Better Work?

Faster work only helps if someone understands it.

But each new tool lets us make more, and each one asks us to think less while we do it.

My daughter goes to a Waldorf school, where the kids make their own textbooks. They draw the pictures and write stories and poems about what they’re learning. By the end, they know the material because they made every page.

Writing by hand works the same way on a smaller scale. An A and a G take two different movements of the hand. Typing turns every letter into the same press of a key. 

AI goes further. You type one request, and three pages come back in the time it used to take to write one good paragraph.

Jerry Seinfeld still writes his jokes by hand on a legal pad. When this came up with my team recently, I put it this way: He’s not cutting corners. He’s not even typing. He needs to know every word.

Some research backs this up, but it’s early. In a 2024 EEG study in Frontiers in Psychology, Van der Weel and Van der Meer recorded brain activity in 36 university students as they wrote words by hand and typed them. Handwriting produced far wider connectivity across the brain.

Typing has a mixed record. In a well-known 2014 study, Mueller and Oppenheimer found that students who typed their notes learned less than students who wrote them by hand. When Morehead, Dunlosky, and Rawson repeated the study in 2019, the groups did not consistently differ. Their combined analysis found only a small, nonsignificant edge for handwriting.

So I won’t say typing makes you forget, but early studies suggest we also pay less attention to the work as it gets easier.

What Does the Research Say About AI and Ownership?

Early studies suggest that people who use AI to produce work feel less ownership of it, and that people who trust AI most think least critically about its output.

In 2025, a team at the MIT Media Lab had 54 people write essays over four months, using ChatGPT, a search engine, or no tools at all. The ChatGPT group showed the weakest brain connectivity. They reported the least ownership of their essays and struggled to quote what they had just written. One caveat: the study is a preprint and has not been peer reviewed, and only 18 people completed its final session.

Now the last session was the most interesting for me: People who had written without tools switched to ChatGPT, and they showed higher memory recall and stronger brain activity. People who went the other way, from ChatGPT to writing on their own, showed signs of under-engagement. 

So, the order was the differentiator. Thinking it through first, then using the tool, worked better.

Something similar shows up at work. A 2025 study from Microsoft Research and Carnegie Mellon surveyed 319 knowledge workers about 936 real tasks they did with AI. Workers with more confidence in AI reported less critical thinking. Workers with more confidence in their own ability reported more. The study is peer reviewed, but it measures what people say about themselves, not what they did.

Lastly, in a field experiment with 758 consultants at Boston Consulting Group, published in Organization Science in 2026, consultants using AI on suitable tasks completed 12.2 percent more tasks, finished 25.1 percent faster, and produced work rated more than 40 percent higher in quality. On one task chosen to fall outside what AI did well, consultants using AI were 19 percentage points less likely to get the right answer.

To me, that’s not a reason to avoid AI. It’s a reason to use it with good judgment. AI is great at certain tasks, we just have to know which ones.

What Does Output Without Ownership Look Like?

A friend of mine calls it a slop flop. He ran a set of customer interviews himself, then used AI to work through the transcripts and build a presentation for the client. He knew the client. He had done the interviews. He believed he owned the material.

Then he got in front of the client and froze. The presentation was good, but he hadn’t worked through all of it himself, and it showed.

A CEO I know saw something similar. One of his junior employees noticed their paycheck was $100 short. The normal move is to tell HR. Instead, the employee asked AI, which drafted a threatening email citing Department of Labor rules.

AI is so good at articulating any case. If you don’t have experience and context, it can lead you the wrong way. I’ve done this too. I’ve gone down a wormhole with AI and ended up thinking the way it did instead of my own way.

It’s the same problem behind a lot of offshore outsourcing pitfalls: you go fast now and fix mistakes later. As I said to my team, AI “slowly strips away the context of the business.” If that goes on long enough, you’ve “evaporated the unique thing that makes us critical to that function.”

What Happened When I Tried to Rebuild My Company Around AI?

It didn’t work. People still did the work.

LinkedIn and X were full of posts about bringing AI into every workflow. I got seduced by all that. I also developed some anxiety and fear that this staffing company was about to be obsolete.

So I started to think about More Staffing as operational leverage instead of human horsepower. I mapped a path from placed talent, to intelligent workflows, to AI agents. I modeled it with an LLM, and it kept patting me on the back. I had built my own echo chamber.

Then I built an app for a business I work with. It caused a lot of stress and took an enormous amount of time. I had to make sure it was accurate, get the team to use it, and teach them how. I couldn’t even prove it would have helped the business. In the end, it went unused.

People and relationships still trump everything. A company can only take on new tools as fast as its people can learn them.

It’s the same problem as the slop flop, just bigger. A person can produce work faster than they can understand it. So can a company.

There is a narrative out there that the world is participating in the illusion of productivity. I agree. All productivity is not equal, and a lot of it is a waste of time.

Now I follow one rule. If it doesn’t grow revenue, if it doesn’t increase profit, don’t do it. The result shows up on your income statement, not on your token usage.

To be fair, there’s another side. Large companies can absorb a lot of tokens and experiment, and some of those tests will pay off. Most businesses don’t need to be at the front of the curve. Being a few months behind costs little when the tools keep getting easier.

AI vs Hiring: Can AI Replace the Hire?

For some work, yes. For work where someone needs to own the result, our clients keep finding that it can’t.

First, the case for AI. If a task is repeatable, clearly defined, and easy to check, give it to AI first. Don’t hire a person to do what a tool already does well.

An apparel brand we work with hired an operations and logistics coordinator through us in June. The role covered purchase orders, freight, ERP updates, and 3PL coordination, so one of the founders could step back from day-to-day operations.

In August, they came back for a second role, an ecommerce operations manager. Before they called us, they had tried to handle that workload in-house with AI tools. It didn’t work. That was their stated reason for coming back to hire a person.

How Should You Use AI on Your Team?

Give it to people who already own the work.

The simplest way to use it is to supercharge employees. They remain responsible for the output, the fidelity of the data, the business objectives, the KPIs, the results. If AI lets them do something quicker or better, go ahead.

What you can’t do is make AI the owner. I said it to my team this way: “We’re these nodes of context still, that require direction.” A tool can draft the work. A person who knows the business decides whether the draft is right. 

This is true whether your team is in your office or overseas. You direct the work and own the results and the KPIs. Each person on your team owns their part of it, and AI is a tool in their hands. The basics of managing offshore teams don’t change because the tools got faster.

This is also why I don’t hire for speed. A fast hire who doesn’t own the work just gives you more to check.

I asked my team a question I still think about: “How much productivity is valuable productivity, if it supersedes what your mind, in a position of ownership, can control?” I don’t have a clear answer. But I know that work nobody owns isn’t real productivity.

Build a Team That Owns the Work

This staffing company is focused on finding excellent people, because I don’t think people are going away.

At More Staffing, we help companies hire vetted Filipino professionals who take ownership of their work. As an offshore staffing partner, we handle hiring, payroll, compliance, and the support that keeps each person engaged.

If you’re weighing AI against a new hire for a role, start your hiring request and we’ll talk it through.