ADVERTISEMENT

Pitch-perfect revenge

AI is turning your breakup texts into bona fide bangers

Illustration: The Telegraph / Mathures Paul

Mathures Paul
Published 10.05.26, 07:55 AM

In the end she packed her bags. He watched. There was nothing to say that hadn’t already been said by not saying anything for two years. A few weeks later, he came across a song on YouTube.

You were so good at starting conversations

ADVERTISEMENT

Turns out you were better at being done

Read at 11:47

Left me on delivered, left me on read

I used to be your favourite notification

What follows are messages they once exchanged, playing out to a riff that would make Barbra Streisand proud:

Never mind,

Forget it

It’s fine

I’ll figure it out

Right, my mistake.

Revenge, it appears, is best served when messages exchanged between ex-lovers become songs. At a time when artificial intelligence is taking jobs and offering questionable advice, the technology has something new to toy around with.

The track Pick Up is based on “pick the f**k up” texts received by London-based Amber Ahn, according to The Times, London, and was subsequently re-recorded by Hackney rapper Not3s. Amber has signed with Broke Records, and Not3s has recorded it in the studio, adding an extra verse written by himself.

Over the past few days, there have been over 4.38 million searches for how to turn text into songs. There has also been a 721 per cent increase in searches for how to make an AI song from texts.

The trend mirros a shift in how AI is being used in music, raising fears that royalties for artistes will be eroded over time by the proliferation of AI-generated sounds. Recently, Taylor Swift’s management company, TAS Rights Management, filed trademark applications to help protect her voice and image from potential AI misuse. This comes at a time when AI-generated artistes such as Sienna Rose are increasingly blurring the line between what is real and what is synthetic in the music industry.

Turning private thoughts and letters into songs has long been a handy tool for musicians. British singer Gwyneth Herbert released the album Letters I Haven’t Written in 2018, prompted by the suicide of a close friend. At first, she wrote letters as a form of catharsis — a creative impulse that took a rather different turn from Ahn’s.

Red Hot Chili Peppers singer Anthony Kiedis reportedly began a relationship with the late Irish singer Sinead O’Connor in 1990, a claim she refuted. Yet the 1991 song I Could Have Lied grew out of an answerphone message O’Connor left, bringing their “relationship” to a close.

Halsey’s 3AM ends with a personal voicemail message: “Your best song is a song that’s currently on the radio. How many people can say that? That their best song is the one that’s currently about to be a massive hit?” The voice belongs to John Mayer. Telephone Line by Electric Light Orchestra is built around the despair of calling someone and getting no answer, only a machine.

Snippets of voice notes Adele recorded while speaking to her son Angelo filter through in the song My Little Love. And there is, of course, Hello, This is Joanie by Paul Evans, a 1970s classic in which the entire song is framed as a voicemail greeting.

Tracks like Pick Up push the limits of such songs further still, pairing the audio with a video assembled on an app such as CapCut. In the last two years, nearly 100 million people have made music on Suno, many of them paid subscribers. The platform is expected to close a Series D funding round in the coming weeks, according to a recent Billboard report, in a deal that could value the start-up at more than $5 billion. Once tracks are created, platforms such as TuneCore are used to distribute them across streaming services.

Suno allows users to create tunes by entering a few words as a written prompt. Select a handful of text descriptors — say, 80s synth pop with dark romance — and an AI-produced song emerges within minutes. Suno claims it trained its model on the basis of fair use, though tech analysts argue the training drew on musicians’ intellectual property without transparency.

Technology has not been kind to the music industry in many ways. The rise of digital music in the late 1990s was met with fierce resistance from music labels. They sued file-sharing platform Napster, which had enabled widespread piracy. It took years to recover, with the likes of iTunes eventually coming to the rescue, followed by Spotify.

Until then, these had been regarded as problems on the business side. Now technology has begun to encroach on creativity itself. Cat Burns, pop star and contestant on the BBC reality programme The Celebrity Traitors, recently said: “As an artiste I am a bit frightened… these songs sound really good.”

On a lighter note, parents too are turning to Suno, transforming everyday texts from their children into emo tracks — capturing everything from demands for Starbucks after school to dramatic insistences that they are absolutely starving.

Privacy experts would counsel caution when it comes to AI prompts. Whatever you upload will help train the models. Chatbots are hungry for input, from wherever they can get it, including messages you might consider perfectly innocuous.

Artificial Intelligence (AI) Breakup
Follow us on:
ADVERTISEMENT