Sciences can be seen according to their role in society: explaining natural phenomena, or proposing tools for improving some course of events. The Observatorium will be mentioned as an example of the explanatory perspective in the context of the dynamics of public opinion. Otherwise, examples of sciences for action include economics, as well as engineering or medicine. The climate change movement illustrates science specifically for political action. Other, not so obvious example of science for action, are the recent advances in AI, known as language models that are based on human data. When presenting human values or political opinions, data proxies based on such models may behave as recursive algorithms and influence structural values of humanity, namely endangering diversity. This feature will be suggested for discussion.

--