Three and a half years ago, I started using AI out of curiosity.
At first, my expectations were modest. I worked with communication materials, reports, and various forms of qualitative data, so I experimented with classification, summarization, editing, and translation. The interaction was largely transactional: I provided material, and the AI helped me process it more efficiently.
My main question was simple: Can this technology actually do useful work?
The answer turned out to be yes. But looking back, I realize that the most significant change has not been in the technology itself. It has been in how I learned to use it.
One important observation is that this journey has been cumulative. Each new use has been added on top of the previous ones. I still use AI for classification, summarization, drafting, and content creation. Those early use cases never disappeared. Instead, they became part of a broader and more layered relationship.
AI as an Analyst
As my confidence grew, I started bringing more complex challenges into the conversation.
AI helped me make sense of large collections of qualitative material: open-ended survey responses, workshop outputs, sticky-note exercises, interview notes, and other forms of unstructured data. It also became useful when preparing leadership workshops, training sessions, consulting assignments, and difficult communication situations. Sometimes I would use it to simulate reactions from different stakeholders or test alternative approaches before entering a sensitive discussion.
The question shifted from „What does this material say?“ to „What is really happening here?“
At this stage, AI became less of a productivity tool and more of an analytical partner. Its value often lay not in producing answers but in helping me discover patterns and perspectives that I had not yet considered.
What interested me most was not content generation but sense-making.
From Answers to Thinking
A deeper transformation began when I stopped bringing only problems and started bringing uncertainty.
Instead of asking for solutions, I found myself exploring unfinished thoughts, half-formed ideas, and situations where I did not yet know what I thought. Leadership often involves ambiguity, competing values, and decisions that cannot be reduced to data or procedures. Those became increasingly common topics of conversation.
The question changed again.
Not „What should I do?“ but „How should I think about this?“
In the spring of 2025, I gave my AI assistant a name: vEikko.
For me, this was not about pretending that the AI was a person. Rather, it reflected a change in the nature of the interaction. The name created continuity. Individual conversations became part of a longer story. Shared context accumulated over time, and I became more comfortable bringing incomplete ideas into the discussion.
The AI did not become more human.
But the relationship became more coherent.
As the shared context grew, something else changed. The quality of the interaction depended less on crafting perfect prompts and more on maintaining an ongoing conversation. Increasingly, I could start with a brief observation or an unfinished thought, trusting that much of the background was already understood.
This experience has made me somewhat skeptical of the popular idea that effective AI use is mainly about writing better prompts. Prompts matter, especially in the beginning. But over time, I found that shared context mattered even more.
Many conversations gradually moved beyond their original topics. A discussion about communication became a discussion about decision-making. A leadership challenge became a conversation about power, emotions, or organizational culture. Questions about AI itself often led to reflections on learning, cognition, and what it means to think.
Without consciously planning it, I found myself engaging in increasingly meta-cognitive conversations.
The Question of Ownership
This journey has not been entirely comfortable.
Alongside the benefits, there has been a recurring concern: what happens to my own thinking when I increasingly rely on a system that can analyze, summarize, and generate ideas at remarkable speed?
The question is not unique to AI. Throughout history, new tools have changed how humans think and work. Writing reduced the need for memorization. Calculators transformed arithmetic. Navigation systems altered our relationship with orientation and place.
AI raises similar questions.
At times, I have wondered whether some intellectual abilities might weaken through disuse. Whether convenience might gradually replace effort. Whether reflection itself could become outsourced.
What I have learned, however, is that the crucial issue is not whether AI participates in thinking. The crucial issue is whether ownership remains clear.
Over time, my concern shifted from dependency to responsibility.
The more consciously I reflected on the relationship, the clearer the boundaries became. AI could help me explore ideas, challenge assumptions, and reveal patterns. But responsibility for judgment, meaning, values, and decisions remained my own.
The tension never disappeared entirely. Perhaps it never should.
Today, however, it feels less like a threat and more like an ongoing discipline: learning how to use AI extensively without surrendering authorship of one’s own thinking.
What I Learned
When people discuss artificial intelligence, they often focus on how the technology is evolving. My experience suggests that an equally important story is how users evolve.
I began by asking whether AI could help me do things more efficiently.
Then I asked whether it could help me understand complex situations.
Eventually, I began asking whether it could help me understand my own thinking.
The irony is that some of the most common public debates about AI now feel less important to me than they once did. Questions such as whether AI is „just“ predicting words are technically interesting, but they do not fully capture the lived experience of working with these systems over time.
Human beings are also prediction machines in many ways. We constantly anticipate, interpret, associate, and revise our understanding of the world. The more interesting question is not how the mechanism works in isolation, but what emerges when it becomes part of a human thinking process.
That was not the journey I expected to take.
And I suspect I am still only at the beginning.
AI can participate in my thinking, but it cannot own it.
Like many of the insights described above, this text was not written alone. It emerged through a dialogue between a human reflecting on three and a half years of AI use and the AI companion that has been part of that journey. The thoughts are mine; the conversation that helped shape them was shared.
Sami Kallioinen is Communications Director of the Tampere Parish Union in the Evangelical Lutheran Church of Finland. He shared the story how he is using AI at the European Christian Internet Conference in Rome whose main focus was AI Usage.
