The following text is a brief commentary on Michael Ochsner's article “The Mis‐conceptualisation of Societal Impact: Why the Swiss Approach to Societal Impact is Productive and not Inexistent” in this special issue. It makes three points about the concept of social impact. The first point questions the appropriateness of the concept of social impact for assessing the value of research to society. The second point concerns the balance between creativity and research limits and between research freedom and responsibility to society. Finally, I look at the implications of information and communication technologies and artificial intelligence in assessing the value of research to society.