My voice was utilized by AI to advertise a dreadful cartoon. It was sheer theft: PAUL GAMBACCINI

Music is my life. Great pop – going all the way back to the rock’n’roll of my childhood to the heart-pounding dance tracks of today – fills me with joy.

I’ve spent my career sharing that joy with countless millions of listeners. But where’s the joy in synthetic pseudo-pop, auto-generated by computers?

Soulless, fake, derivative and artificial, it lacks everything that makes music magical. Worse still, AI is ripping off all the great musicians, past and present, who mean so much to me – while ensuring that the next generation is starved of inspiration.

Most musicians are not stars. But if they’re lucky, they can earn a steady living behind the scenes, working for hire in studios and on soundtracks.

Their livelihoods are set to be wiped out as AI music replaces them. It won’t be nearly as good. But it will be cheaper, and infinitely more convenient.

The Big Tech companies responsible are training their AI systems to analyse all the recorded music in the world, with special emphasis on what is current – a process known as ‘scraping’.

According to a global study published last December, the cost of AI to the music industry will be $10 billion (£7.9 billion) over the next five years.

But that’s just the start. By the end of the decade, tech giants will pocket around $48 billion (£38 bn) by making music with AI, according to the study by CISAC, a French organisation that campaigns on behalf of artists.

AI is ripping off all the great musicians past and present who mean so much to me – while ensuring the next generation is starved of inspiration, writes Paul Gambaccini

We need those artists. A UK music industry report last November showed British artists are contributing more to the economy than ever before – a record £7.6 billion last year, up by 13 per cent from £6.7bn in 2022.

Our music exports revenue is also at a record high of £4.6 bn, up 15 per cent from £4 bn two years ago. And 216,000 people work full-time in the industry.

Billie Eilish, Jon Bon Jovi, Katie Perry and Miranda Lambert are among the stars who have spoken out against AI stealing their music. In fact, I don’t know a single artist who is in favour of it. And I’ve already been a victim myself. A friend, almost incredulous, flagged up a YouTube video to me. ‘This isn’t really you, is it?’ he said.

It was a promo for an animated movie, a dreadful piece of rubbish featuring the Three Musketeers as dogs. Blue-tinged fluffy dogs. They looked like pets for the Smurfs, drawn by a child of limited artistic talents.

And over the top was my voice, proclaiming it to be the most wonderful and hilarious thing I’d ever seen. My accent, my delivery, my way of using pauses and inflexions and rhythm, were being perfectly mimicked by AI. It was sheer theft.

I’ve been reviewing movies for about 15 years, and I struggle to think of anything I’ve watched that was as bad as this. But no one who happened to click on this YouTube trailer would have any reason to doubt it was me, voicing my genuine opinion.

Even to my own ears, it sounded like my voice.

And that’s the hidden danger of AI. These programmes are so highly evolved that they’re easily capable of creating infinite numbers of fakes. In fact, accurate versions of an individual human voice can be generated from just 30 seconds or so of audio.

My voice is the most valuable attribute I possess. And that’s true of all the great DJs and presenters I’ve worked with. Listeners can recognise them with just a couple of words.

With AI, any celebrity’s vocal tones could be used to front commercials for products they would never use; or to sing songs that are not their own.

The most notorious digital rip-off is a track called Heart On My Sleeve, which landed on TikTok and Spotify two years ago.

On a first hearing, it appeared to be a collaboration by two of the world’s biggest acts, Drake and The Weeknd.

But it lacked feeling, and the wit that is the trademark of both artists. And the production was terrible – amateur and scratchy. It was fake. A TikTok user under the handle Ghostwriter977 had apparently fed an AI programme with everything ever recorded by rapper Drake, who has sold more digital singles than any other performer, and The Weeknd, whose hit Blinding Lights was on the Billboard Hot 100 for a record 90 weeks.

Within a week, Heart On My Sleeve was viewed more than nine million times on social media, before it was removed in response to a legal challenge by the artists’ record label, Universal.

On My Sleeve was released by an online content creator called Ghostwriter, who used AI to make it sound as if it were sung by Drake (left) and The Weeknd (right)

You might imagine that it’s completely illegal to copy the unmistakable voices of two such world-famous performers. But in other cases, it’s been difficult to prove which laws are being broken or bent.

Heart On My Sleeve was generated by mimicking the artists’ sound. But it also drew on similar music which has its roots in the pop, soul and R&B of the past 60 years.

One US lawyer specialising in intellectual property rights, Kate Downing, summed it up neatly in the New Yorker magazine in 2023: ‘Mathematically speaking, the work comes from everything. It may well be argued that the use of any one image from the training data is not substantial enough to call the output a derivative work.’

In other words, if it’s copying everything, then legally it’s copying nothing.

I related the story of me and those sickening dogs to my dear friend Sir Stephen Fry over lunch last week. The implications horrified him. Future generations of musicians and artists, he predicted, will be deprived of any just rewards for their ingenuity.

Every thought you ever have can be stolen by AI; and every thought that you have not had could be presented as if it was your own. And the only possible defence that we have against this – whether as artists or society as a whole – is strong legislation.

It’s unthinkable that the Government should legitimise these practices by handing tech giants a licence to take whatever they want and reproduce it with AI.

Never mind fake news. Fake voices are an even greater danger.