Conversational AI as a tool for thinking-with rather than a technique for replacing our labour

What I found fascinating about Twitter in the olden days was how readily it encouraged the development of ideas. The brevity of the format and the always-on character of the platform meant it always provided an occasion (in the dual sense of ‘a particular event, or the time at which it takes place’ and ‘reason; cause’) for sharing an idea. Often it would be met by others sharing interlocking ideas in ways which enriched the process. Even when this didn’t happen, the act of sharing it helped give form to something which had previously been nebulous. In exteriorising ideas we give them shape, helping form more concrete epistemic objects in our own minds which we can work with and develop over time. The same is true about blogging, as evidenced by this post in which I’m trying to give shape to a latent idea I’ve had itching at the back of my mind for years. It matters that the occasion comes from ‘out there’ and is at least in principle social. Obviously ideas can be exteriorised through private note taking but this is not occasioned in the same way, creating a different dynamic of exteriorisation. For one thing it is entirely self-generating, rather than resting on the mobilisation of desire which takes place through social platforms. The intersection of social possibilities and algorithmic incentives can be profoundly generative, even if the balance has turned in a direction in the former is increasingly consumed by the latter.

My suggestion is that conversational agents represent occasions for exteriorisation of ideas in the same way. To seize on what C Wright Mills called ‘the feel of an idea’ and elaborate upon it is a deeply satisfying process. Furthermore, it is one which by its nature is iterative and tends towards incompletion. It’s less a matter of outcomes and more about concretising epistemic objects in order that it becomes possible to do more expansive and practical things with them i.e. write books about them, build interventions which leverage them, make recommendations to others. The exterioisation of ideas is how we furnish our intellectual world, building on but never reducible to the things we read and the conversations we have (the interiorisation). This is how we should understand the scholarly use of generative AI, as a tool for thinking-with rather than a technique for replacing our labour.