Is the emergence of artificial subjectivities endowed with a certain autonomy compatible with Habermas’ theory of discourse? To answer this question, it is necessary, preliminarily, to address the question of the Habermasian subject. In doing so, one observes how machines stand as potential interlocutors. As interlocutors, like the other protagonists in Habermasian discourse, they must aim at understanding and must have a legitimate interest. This opens the field to two orders of problems: (a) are machines capable of understanding? (b) can machines have an interest? Arguments borrowed from the Chinese Room debate show how, in principle, there are no obstacles to the entry of machines into discourse. If machines can participate in discourse, then in what language do they talk to humans? To communicate, humans and machines must resort to translation techniques, known as natural language processing. Translation already places itself outside language, deferring to a shared idea of reality in which to root universal meanings. By placing communication outside language, the entry of machines into discourse seems to checkmate Habermasian theory. However, two further arguments – “the Others mind reply” and the private language argument – make the case that communication between humans and machines poses no different problems than communication between humans alone.
