This paper shows that generative artificial intelligence is changing how the past is revealed to humans and how humans remember the past by reshaping the ethics of collective remembering and forgetting. Artificial intelligence is changing the nature of collective memory in a process that turns history from an object of media representation into an object of algorithmic performativity. By relying on the framework of postphenomenology, the analysis shows that artificial intelligence makes humans engage with notions of togetherness and boundlessness, gradually substituting issues of authenticity and accuracy as main referential frames in historical knowledge production. Consequentially, artificial intelligence mediates a dialogic version of historical awareness, which makes the past responsive to the distributed actions of human and non-human assemblages. In this process, generative artificial intelligence redefines what it means to be responsible and accountable in preserving, transmitting and promoting historical legacy by putting to work new mnemotechnic values.