Artificial intelligence is radically transforming the landscape of contemporary poetic creation, giving rise to new expressive forms such as generative videopoetry and augmented poetry slam. Poetry and poetic languages hybridize with intelligent technologies toward a relational, participatory, and multimedia aesthetic, where poetry itself is understood as a fluid process traversing pre-text, intro-text, sub-text, super-text, and post-text. Recent studies show that poetry generated by language models is now often indistinguishable from human poetry for non-expert readers, receiving higher ratings in terms of rhythm and beauty. This situation requires rethinking AI's role not as a mere tool, but as a cultural agent and generative environment, and questioning the ethical conditions of its integration, also in light of the new European AI Act norms, which introduce transparency obligations on training data for general-purpose models. The article argues that these hybrid practices produce a new aesthetic - characterized by procedural fluidity, medial hybridization, and negotiation between human and machine - that manifests exemplarily in generative videopoetry and augmented slam, requiring a reconsideration of traditional criteria of authenticity, authorship, and responsibility.
Dimitri Ruggeri (Sun,) studied this question.