From Using AI to Working With AI

For most of my professional life, technology was something I used. Software helped me communicate, organize information, manage projects, build businesses, and make decisions. Even powerful digital tools remained tools: I gave an instruction, received an output, and decided what to do next.

Artificial intelligence changed that relationship.

My experience working with ChatGPT, Claude, Gemini, and Codex has gradually changed the way I think about professional work itself. AI is no longer simply a place where I ask a question or generate a piece of text. Increasingly, I work with AI systems throughout the day—as collaborators for thinking, research, planning, building, reviewing, and solving problems.

That distinction—between using AI and working with AI—has become one of the most important changes in my professional life.

It Started With Questions

Like many people, my early experience with generative AI was conversational.

I asked questions. I explored ideas. I requested explanations, summaries, drafts, and suggestions. The immediate value was obvious: information that might previously have required considerable searching and synthesis could often be organized much faster.

But the more I worked with AI, the more I realized that asking isolated questions was only a small part of what was becoming possible.

The real change began when conversations became continuous work.

Instead of asking AI to complete one disconnected task, I began providing context about larger objectives, previous decisions, constraints, technical requirements, business considerations, and work already completed.

The quality of the collaboration changed dramatically.

AI became more useful when it understood not only what I was asking, but also why I was asking it and what had already happened.

Different AI Systems Became Different Collaborators

Working with several AI systems also taught me that I should not think about artificial intelligence as one interchangeable capability.

My experiences with ChatGPT, Claude, Gemini, and Codex have been different.

I use them in different ways depending on the work in front of me.

Sometimes I need to explore an idea and challenge my own assumptions. Sometimes I need extensive technical implementation. Sometimes I need research or another perspective. Sometimes I want an independent review of work that another AI system has already completed.

That last point became especially important.

If one AI system builds something, another can examine it.

That does not eliminate the need for human judgment. It creates another layer of review.

Over time, my workflow began to resemble something less like a person operating a collection of software tools and more like a person coordinating different digital contributors.

My Role Began to Change

One of the biggest surprises was that AI did not simply make individual tasks faster.

It changed what I spent my time doing.

When AI can help research an issue, analyze alternatives, draft an implementation, test assumptions, review results, and document what happened, the human role begins moving upward.

I spend more time defining objectives.

I spend more time deciding what should and should not be built.

I spend more time examining risk, questioning results, resolving disagreements, establishing boundaries, and determining whether completed work actually serves the larger purpose.

In other words, AI can perform more of the execution, but that makes direction and judgment more important, not less.

This has been particularly noticeable while building GlobSynk.

There are situations where multiple AI systems contribute to the same body of work in different roles. One may help reason through architecture. Another may implement a bounded piece of work. Another may review or audit it. I remain responsible for deciding what moves forward.

That is fundamentally different from simply asking a chatbot for an answer.

AI Expanded What I Could Attempt

Perhaps the biggest professional impact has been a change in scale.

There are projects I would once have considered unrealistic for an individual founder to coordinate without a much larger immediate organization.

AI changed that calculation.

It did not make difficult work effortless. Complex systems still require planning, iteration, verification, security, governance, and countless decisions.

But AI dramatically increased the amount of intellectual and technical work that I could organize at one time.

That changed my perception of what was possible.

The important lesson for me was not simply that AI could help me work faster.

It was that AI could increase the scope of work I was capable of directing.

Those are very different things.

Speed Is Valuable, but Verification Matters More

Working extensively with AI has also made me more aware of its limitations.

AI can be impressive and still be wrong.

It can misunderstand an instruction, overlook an important condition, make an assumption that was never authorized, or produce something that looks convincing without being sufficiently verified.

The faster AI allows work to move, the more important verification becomes.

That is why I increasingly value workflows where work can be reviewed independently, tested against explicit requirements, and stopped when the evidence does not support moving forward.

AI should not receive trust merely because its output sounds confident.

Trust should come from evidence.

That principle has become increasingly important in the way I work.

The Human Still Has to Decide

There is a temptation to describe the future of AI as a competition between humans and machines.

My own experience has led me toward a different question:

What happens when people become very good at working with intelligent digital systems?

AI can generate possibilities at extraordinary speed.

But someone still has to decide which possibilities matter.

Someone has to determine the objective.

Someone has to establish values and boundaries.

Someone has to decide what risks are acceptable.

Someone has to recognize when a technically correct answer is still the wrong decision.

And someone ultimately has to take responsibility.

For me, that remains the human role.

From Software Tools to Digital Contributors

This experience has also influenced how I think about the future of work.

Traditional software waits for people to operate it.

AI systems are beginning to do something different. They can interpret objectives, work through problems, produce artifacts, interact with other systems, and participate in increasingly complex workflows.

That does not make them human employees.

But it does make the traditional word tool feel increasingly incomplete.

I have found the idea of a digital contributor more useful: a digital system that can contribute meaningful work within boundaries established by people and organizations.

Those contributors may eventually include AI agents, digital workers, specialized assistants, AI clones, and other forms we have not yet defined.

The important question will not simply be what these systems can do.

It will be how responsibly we decide to use them.

AI Changed How I Think About Leadership

Before working this closely with AI, I often thought about leadership primarily in relation to people, organizations, and resources.

I now think increasingly about leadership as the coordination of both human and digital capability.

That requires new skills.

Leaders will need to understand when to delegate work to AI and when not to.

They will need to know when an AI result requires independent review.

They will need systems for identity, authority, accountability, security, and auditability.

And they will need to preserve something that technology cannot decide for them: purpose.

Efficiency is not purpose.

Automation is not purpose.

Intelligence is not purpose.

Those remain human decisions.

What I Have Learned

I am still learning how to work with AI.

The technology is changing too quickly for anyone to reasonably claim that the final model of human-AI work has already been discovered.

But several lessons have become clear to me.

AI becomes more useful when it has meaningful context.

Different AI systems can contribute different strengths.

Independent review is valuable.

Speed without verification creates new risks.

Human judgment becomes more important as execution becomes easier.

And perhaps most importantly, learning to work with AI is different from simply learning how to use AI.

That difference has changed my professional life.

Looking Ahead

I do not believe the most interesting future is one where people disappear from work.

I believe the more consequential future may be one where people learn to coordinate capabilities that previously did not exist.

Humans will continue bringing judgment, responsibility, creativity, relationships, values, lived experience, and purpose.

AI will increasingly contribute analysis, generation, automation, reasoning, and execution.

The organizations that learn how to combine those strengths responsibly may operate very differently from the organizations we know today.

I have already seen the beginning of that change in my own work.

I started by asking AI questions.

Then I started giving it tasks.

Eventually, I realized something more significant had happened.

I was no longer simply using AI. I was learning how to work with it.

J Starr (Dann) is the Founder of GlobSynk. This essay is a first-person Founder perspective for Voices, reflecting his personal experience working with artificial intelligence and does not represent an independent GlobSynk Media assessment of the AI systems mentioned.