I personally believe that artificial intelligence, as a human technology, should be used to flourish human relationships. This is also a thesis expounded on in Ivan Illich’s Tools for Conviviality. A convivial society, Illich claims, is one where “individual freedom [is] realized in personal interdependence”, and because “an individual relates himself… to his society through the use of tools”, so it goes that technology has the capacity to shape human relationships, setting an imperative to assess them.

We are currently in a hype of dreams of weaving large language models (LLMs) into the infrastructure of society, its informative and generative nature lending a tendency to reimagine our communicative infrastructure—and consequently our everyday practices of communication.

This reimagining, often taking the form of “optimising” communication, either by: (1) correction of spelling and grammatical mistakes; (2) translation into a more palatable language (e.g. “explain like I’m five”); (3) summarisation into bullet points; and more, is actually not a novel force of change. Han Byung-Chul calls this transformation “the smooth” in Saving Beauty which was published in 2015(!), remarking that “today, communication also becomes smooth. It is smoothened out into a frictionless exchange of information.” Han characterises the aesthetic of the smooth as the elimination of any negativity, interior, and other, removing any opportunity for judgment, interpretation or reflection. When communication is reduced to smooth information, there is only an encounter with the self without the other, and without the other, there is only the same—society become smooth.

Friction makes communication human because human relationships are built on the friction generated by the turbulence of two interior complexities attempting to consume the other. Friction provides opportunity for action, the “autonomous and creative intercourse among persons” which constitutes Illich’s conviviality. Frictionless communication makes for a diminished social experience, if not the complete absence of it.

Thus, how LLMs will integrate into our society deserves serious contemplation. When we use LLMs to transform our communication, we must be aware of the backwash of how our sent messages are shifting the standard for human communication, how we might pick our words in a new world of communication alongside our artificial interlocutors. How do we nurture human communication in midst of an intermediary AI? Will there be espionage?