Why does using AI still feel uncomfortable?
A remark that stuck with me
While rereading the content of one of my sites, a friend spotted several phrasings typical of generative AI.
He's developed a sharp eye for this, from being exposed to it so often. His reaction was almost visceral: "that phrasing is typical of an AI agent."
That exchange stuck with me.
Not because of the remark itself, but because of what it triggered.
The urge to justify myself
For a brief moment, I caught myself wanting to specify how I had used the AI. What it had written. What I had rewritten. What genuinely came from me.
Why did I feel the need to make that distinction?
The question stayed with me.
I use an AI agent almost every day. To write, structure, proofread. And until that exchange, the question didn't even come up for me anymore.
When I talk about AI in this article, I'm referring to conversational assistants (ChatGPT, Claude...) and the AI agents built on top of them. The term "AI" has become so broad that it deserves clarifying.
A discomfort more widespread than it seems
That moment also reminded me of something I regularly observe around me: many people still hesitate to say they use AI, both at work and in their exchanges (emails, messages).
Around me, I see gradual adoption, often held back by contradictory discourse and terminology that can make these tools feel like they're not meant for everyone.
Setting boundaries is not passing judgment
This reaction doesn't appear out of nowhere. It fits into a context where AI is very often discussed through the lens of risk.
I understand that some organizations regulate the use of AI. Confidentiality, compliance, data protection. This caution seems legitimate to me.
The problem isn't the regulation itself.
The problem arises when this regulation is experienced as a value judgment on the people who use these tools.
A rule meant to protect data or govern usage can easily be read as a signal: "using AI is a practice to be wary of."
But by communicating almost exclusively through the lens of risk, we build lasting distrust. A distrust that eventually settles into our own relationship with the tool.
Nothing to feel uncomfortable about
Using AI is not admitting a weakness. It's choosing a tool when it brings value.
I'd feel more uncomfortable giving up a tool that improves my work, just out of fear of what others might think.
We're probably in a period of transition. A period where we talk more about the tool than about what it actually allows us to produce.
Maybe the real sign of maturity won't be when everyone uses AI.
It will be the day their use no longer triggers surprise, or justification.