Writing
AI and Life as We Know It
This essay preserves the perspective of early 2023, when I first explored ChatGPT. Its expectations about the future are presented as questions and possibilities, rather than settled forecasts.
When I first used ChatGPT, what struck me was the range of things I could attempt within one conversation. I could explore a movie plot, generate code, draft a blog, plan a trip to Scotland, and work through ideas for a fitness routine.
The outputs still needed judgment. A plausible answer was not necessarily a correct one. But the experience changed my sense of what it meant to begin a task.
Something I would once have approached through a long search, a specialist tool, or a blank page could now start with a question in ordinary language.
OpenAI had introduced ChatGPT on 30 November 2022. For me, that release made AI feel immediate and personal. It was no longer only a technology I read about. It was something I could work with.
What happens to cognitive labour?
Much of a working day involves preparing to do things. We organise information, draft presentations, compare options, write messages, search for explanations, and decide what needs attention next.
This is cognitive labour. It can be valuable, but it can also consume the energy we hoped to spend on the work that matters most.
AI made me wonder how much of that effort could change. What if a first draft took minutes instead of hours? What if a programmer could explore several approaches without writing every line from scratch? What if learning a new subject began with an explanation tailored to the question actually confusing us?
The possibilities were exciting. They also raised a second question: what would we do with the capacity we gained?
We might use it to think more carefully, rest, learn, or spend time with people. We might simply fill it with more tasks. Efficiency does not choose between those futures for us.
A draft is not the whole job
My first reaction was to imagine entire professions disappearing. If a system could write code or an essay quickly, what would happen to developers and writers?
But producing an output is only part of professional work. Someone still has to understand the problem, identify errors, judge consequences, and take responsibility for the result.
A generated application is not automatically a reliable product. A fluent explanation is not automatically knowledge. A diet suggestion does not turn the person requesting it into a nutrition specialist.
These distinctions matter because access to a capability can feel like mastery of it. AI may lower the cost of attempting work without removing the need to understand what good work requires.
That still has enormous implications. Some tasks may become less valuable in isolation. Others may become more important: framing a problem, connecting different kinds of knowledge, testing an answer, and deciding which result deserves to be used.
I do not know exactly how that will reshape employment. I do think it asks us to examine what part of our expertise is routine execution and what part is judgment developed through experience.
When a title becomes less stable
We often answer “Who are you?” with a profession. Engineer. Writer. Designer. Manager.
There is nothing wrong with taking pride in work. But if a tool changes how that work is done, an identity built entirely around performing the task can become vulnerable too.
What remains when I am no longer the only person who can produce a certain kind of output?
For a writer, perhaps it includes a way of noticing, a lived experience, or a question that feels worth pursuing. For an engineer, it may include the ability to see a system as a whole and recognise what will fail outside the demonstration.
Those qualities also need development. We cannot protect ourselves merely by calling them uniquely human. But they move the conversation beyond the speed at which a task can be completed.
The possibility I found reassuring was that making something easier to produce does not necessarily remove the joy of making it. If I write because I want to understand an experience, a machine's ability to generate sentences does not take that experience away from me.
Intelligence and experience
AI also brought me back to consciousness. A system can produce language about love, grief, beauty, or peace. That does not by itself establish that it experiences any of them.
The broader question of machine consciousness remains open. I cannot settle it by intuition or by the quality of a conversation. What I can distinguish is the output I receive from the experience I have while receiving it.
When I lie on a beach, listen to music, or miss someone, there is a lived dimension to the moment. It matters to me beyond anything I could describe or produce from it.
That distinction made me question how often I had treated my own mind chiefly as a machine for getting things done.
If more cognitive work can be assisted, perhaps we will have to become more deliberate about the ends it serves. What is all this capability for? What do we want our days to feel like? What deserves our attention once producing more is no longer an adequate answer?
The hope beyond efficiency
I do not expect technology to make us wise automatically. It can amplify distraction as easily as it can support inquiry. It can free time or create new demands on it.
My hope is that AI will make a neglected question harder to avoid: who are we when we are not occupied with proving our usefulness?
That is where this technological change meets the inward inquiry I keep returning to. We may gain extraordinary tools for producing, planning, and solving. We will still need to decide what kind of life those tools should help us build.
The future interests me for that reason. Alongside the question of what AI will do, I want to keep asking what we will choose to do with ourselves.
Reference: Introducing ChatGPT — OpenAI, 30 November 2022.