Music and Ia

How artificial intelligence is reshaping pop in Italia

Mina singing Mahmood and videos generated out of thin air: are we witnessing the birth of a new aesthetic or the end of music as we have always known it?

by Gabriele Amante

13 February 2026

4' min read

Translated by AI

Versione italiana

4' min read

Translated by AI

Versione italiana

With the album 'Dilettvoli eccedenze 2', released at the end of 2023, Mina released the track 'Abban-dono': a pop song with acid notes, poised between swing and jazz. The video clip, launched on the singer's official channel on 5 December 2023, was made by IULM AI Lab, the artificial intelligence laboratory of Milan's IULM University.

The video, which aims to turn Mina's career and life into a sort of virtual museum, opens with the image of a white tiger with laboured breathing in Piazza del Comune in Cremona. The reference needs no explanation, immediately evoking the nickname the singer has carried since the beginning of her career. Picking up on the concept of the museum, during the video we see several female visitors - in the likeness of the singer - examining paintings from various painters and eras, such as Botticelli's Venus or the Mona Lisa.

However, watching the video today, one can clearly perceive the weight and limitations that AI had in 2023: people or objects never have a precise shape, they change frame after frame and have very bright colours. Yet from the perspective of an artist who retired to private life decades ago but still successfully pursues her career, the idea remains exceptional. Thanks to AI, Mina is heard but, finally, also seen. It is not her physically, of course, but the artificial intelligence gives us an almost real Mina, going beyond the concept of mere animation or the retrieval of old repertoire clips. It is a Mina who embraces innovation to be present not only with her voice, but also with her image.

Deepfakes and Viral Covers: When AI 'steals' the voice

But in addition to video clips, a relevant and still topical phenomenon is that of AI covers. Through digital models called RVC (Retrieval-based Voice Conversion), specially trained to imitate the voice of the 'Tiger of Cremona', it is possible to make her interpret the hits of the moment. During the Sanremo 2024 period, numerous artificially generated versions of the singer performing the songs in the Sanremo competition were all the rage on social media. 'Tuta Gold' was one of the most clicked tracks in this 'miniaturised' guise.

These are songs that never make it onto official streaming platforms, but are a huge hit on social media. Often the result is barely discreet: the feeling is that of a simple change of timbre of the original singer, with no real artistic contribution to justify listening. Other times, however, the result is excellent.

A noteworthy experiment, which deserves a careful listen, is the AI cover of Nayt's 'Se ne va', 'sung' by Fabrizio De André. In this case, the algorithm perfectly matches the atmosphere of the original: the very low key and the constant arpeggio of the guitar are fully in Faber's vein, recalling with incredible fidelity the first part of the Genoese singer-songwriter's production.

The Underdog Case

Well before all this, there was the case of the song 'Cold Moon In Deep Water'. The Underdog band turned to the Academy of Arts and New Technologies (AANT) in Rome for the production of what was one of the first Italian video clips made entirely with artificial intelligence.

Being a video produced almost three years ago now, it is evident how outdated the technology of the time is by today's standards. However, it must be acknowledged that for the genre of the piece - characterised by precise cadences and a rhythm section without cymbals or cymbals, consisting solely of drums - the end result is a perfect combination of music and image. On the one hand, we have an uncertain and almost hypnotic jazz sound, on the other hand, images that change frame after frame, with acid-looking pastel colours that perfectly match the atmosphere of the piece.

AI, an ally for all ages

Artificial intelligence, in the case of an artist like Mina - who has long since decided to retire to private life without giving up music - makes it possible to make her present not only through her voice. A video celebrating her career, comparing her to a museum, manages to give the song a visual body. AI thus fills a gap that many fans have felt for years: it is not a simple animation, but a succession of generated images that offer something new for the singer, going beyond the banal use of archive footage.

However, the case of the Underdog suggests that artificial intelligence has explored niche genres before making its way to mass pop. If this technology works on grunge- or jazz-style songs, where the impact of metamorphosis is powerful, then it can also work on simpler, 'democratic' listening pieces. But what is also democratic is the use of the technology itself: AI can be used in the world of memes, making artists far removed from those lyrics perform unlikely songs, but it can also fulfil the desire to hear how a current song would be sung by a legend of the past.

Let's be clear: the aim is not to legitimise the use of voice without due permission, but to emphasise that we have reached a point where customisation of listening is total.