A few weeks ago, during a conference on artificial intelligence for science, someone in the audience asked a very common question. The framing was familiar because it appears almost systematically whenever new technologies are discussed: “AI is a tool and like any tool, it can be used for good or bad. So how do we make sure it is used responsibly?”
Deep Tech startups are often presented as key actors in the ecological transition. However, the environmental and societal impact of Deep Tech startups is rarely addressed seriously during the early stages of technological development.
For a long time, research evaluation was primarily organized around scientific excellence. The central questions were relatively clear: Is the science robust? Is it novel? Can it advance knowledge? Eventually, additional criteria such as publication metrics, patents, technology transfer became increasingly important. But the overall logic remained largely the same: research produced knowledge first, and societal effects would follow later. This is progressively changing.
Low-tech is often presented as a counter-model to high-tech. Simpler technologies, fewer resources, more resilience: a way to reconcile innovation with ecological constraints. This suggests that the problem lies in the level of technology itself. As if the question were simply to choose between more or less sophisticated systems. It is not.
Autonomous vehicles are currently presented as one of the most promising technological breakthroughs in mobility. Safer roads, fewer emissions, better access to transport, more efficient systems: the list of expected benefits is long, and by now familiar. These promises are not just speculative. They are already structuring industrial strategies and public policies.