I'm skeptical about revolutions — for political revolutions, I doubt they succeed; with technological ones, I doubt they really happen. Sure, LLMs and agentic AI are changing our society. Before them the internet did, and before that television, cars, radio, the telephone, electricity, the steam engine — and we could keep going back at least as far as the invention of the alphabet. But were all these innovations «revolutions» that turned humanity in a new direction? That, after all, is what a revolution is: a change of direction. I suspect most of these inventions changed our speed instead, sometimes with a small jump, sometimes with a giant leap, but kept us going the same way.
So I appreciated it when, answering my question about AI, Thomas Schneider started with the engines that two hundred years ago took over repetitive physical work — just as artificial intelligence is now doing with a growing share of cognitive work. He brought it up to make a point. Nobody ever wrote a single engine law: over time we ended up with thousands of rules — for cars, roads, drivers, airplanes, factories — each shaped by the risks of a particular context. And presumably we will do the same with AI. What is new, in his view, is the speed, and the fact that an algorithm, unlike an engine, can be copied and moved around the world in milliseconds.
Schneider is a Swiss ambassador and co-director of international affairs at the Federal Office of Communications (OFCOM). For more than twenty years he has led Swiss delegations in the main international forums on digital governance, and he chaired the Council of Europe negotiations on the Framework Convention on Artificial Intelligence, the first binding international treaty on AI. He is now among those preparing the Geneva AI Summit, which Switzerland will host on June 21–22, 2027, under the motto «Bridging Innovation and Trust». More details are due tomorrow, November 11, when Federal Councillor Albert Rösti and Fabiola Gianotti, former director-general of CERN and the Federal Council's delegate for the summit, will present the preparations to the press.
I met Schneider ahead of his appearance at a forum on AI in biomedicine organized in Lugano by IBSA Foundation and the Dalle Molle Institute for Artificial Intelligence (IDSIA). His official role put some limits on the conversation: he cannot express opinions, not even personal ones, on matters outside his office's remit — which is why the interview doesn't cover the climate impact of AI.
What follows is a lightly edited transcription of the rest of our conversation.