The Impact of AI on Human Interactions

The genuine productivity benefits of AI personal assistants are worth taking seriously before complicating them, because they are real and in many cases transformative. The administrative burden of modern professional life is genuinely enormous and genuinely costly to human attention. The average knowledge worker spent a significant portion of their working week on tasks that required time and cognitive overhead but added little distinctive value: scheduling coordination, email triage, summary writing, research aggregation, meeting preparation. AI assistants that handle this layer of overhead give people something genuinely scarce and valuable, time and mental space for the work they actually care about.

There is also an accessibility dimension that does not get enough attention. For people with certain disabilities, cognitive differences, or executive function challenges, AI personal assistants are not a productivity optimization. They are an enabling technology that makes professional life navigable in ways it was not before. Someone who struggles with the organizational overhead of managing a complex calendar, synthesizing large amounts of written information, or drafting formal communications that meet professional norms, may find in AI assistance the scaffolding that allows them to bring their genuine capabilities to bear without being defeated by the administrative layer.

At its best, AI assistance does something that all good tools do: it makes the human using it more fully themselves. A musician who does not have to tune their own instrument by ear has more energy for the music. A writer whose research is organized by an AI has more time in the particular quality of attention that writing actually requires. The argument for AI personal assistants, made honestly, is not that they replace what is distinctively human about professional work. It is that they clear away the parts that were never distinctively human to begin with, leaving more room for the parts that are.

Here is the uncomfortable part. A lot of what AI personal assistants are most efficient at handling is the ambient communication that, while it looks like overhead, is actually doing relational work at the same time. When I write a quick email to a colleague checking in on a project, I am not only transferring information. I am also signaling that I am thinking about them, that our collaboration matters to me. When an AI drafts that email, or when I use AI so efficiently that I stop sending those check-ins because the information is captured elsewhere, the relational signal disappears along with the overhead. The colleague gets fewer messages. The project may run just as smoothly. But something in the fabric of the working relationship has changed, and it has changed in a direction that tends to feel like efficiency from the inside and something colder from the outside.

There is research on this, and while it is still early, the direction is worth paying attention to. Studies of remote and hybrid workplaces have found that the relationships that sustain most easily are the ones with enough incidental contact to feel warm, not just the ones that are most productive. The coffee machine conversation that serves no agenda. The email that starts with a line about the weekend before getting to business. The response that takes a moment to acknowledge the difficulty of what the other person is dealing with. These interactions have always been considered slightly inefficient by the standards of pure task completion. They are also the connective tissue of working relationships, and they are exactly the kind of thing that AI assistance, used without awareness, tends to gradually optimize away.

And there is a more personal dimension worth naming. When we delegate more of our communication to AI, we spend less time in the actual practice of putting our thoughts into words for another person. Writing to someone, really writing, is a form of presence. It requires you to think about who they are, what they know, what they need to understand, how they are likely to receive what you are saying. That practice, repeated over a professional lifetime, builds a kind of relational and communicative intelligence that atrophies when it goes unused. The worry is not that AI will write slightly generic emails on our behalf. The worry is that over time, we slowly lose the habit of the real thing.

The answer to this is not to stop using AI personal assistants. That would be a bit like arguing against calculators on the grounds that mental arithmetic builds cognitive strength, technically true and practically beside the point. The answer is to develop a much more deliberate relationship with what we delegate and what we keep, and to be honest with ourselves about the difference between tasks that are genuinely routine and tasks that are doing relational work we might not consciously notice. Scheduling a meeting between two people who already know each other well is genuinely routine. Checking in on a colleague who is navigating a difficult project is not, even when it looks like a three-line email.

There is also a case for actively maintaining some inefficiency as a form of relational investment. Not every interaction needs to be optimized. Not every response needs to arrive in the minimum possible time. The colleague who gets back to you a day later with a response that feels genuinely considered, that references something you said in a way that indicates they actually read and thought about it, is doing something more valuable than the colleague whose AI sends a prompt and well-structured reply within minutes. Slowing down in specific contexts, reserving actual human attention for the communications that warrant it, is not a failure to adopt best practices. It is a choice about what kind of relationships you want to have.

I still use my AI assistant every day. I am not going to pretend the time it saves me is not real. But I think about my grandmother's letters, handwritten, specific, dense with the particular texture of a mind attending to another mind across distance and time, and I notice the distance between that kind of communication and the fluent, slightly frictionless outputs that AI generates. Both are communication. But they are not the same thing, and knowing the difference, and making choices accordingly, seems like one of the more important habits to develop in a world where the efficient option is always available and the human option requires a small act of will.