Josh Fawaz’s Suno Workflow Reignites Debate Over AI in Hitmaking

Josh Fawaz has responded to his critics in perhaps the most revealing way possible: by showing them exactly how generative artificial intelligence fits into his music-making process.

The chart-topping DJ and producer shared a behind-the-scenes video on Instagram this week, captioned, “Behind the scenes on some vocals to my songs.”

In the clip, Fawaz records himself singing into a microphone. The performance is deliberately rough — terrible, by his own admission — before he feeds the recording into Suno, a generative AI platform capable of creating vocals, instrumentation and entire songs.

The resulting vocal bears little resemblance to what Fawaz originally sang. His voice is transformed into a smooth, polished R&B-style performance, complete with a markedly different tone, refined phrasing and an American accent.

It is, to be clear, considerably more than the widely used, industry accepted AutoTune.

Pitch-correction software may adjust the notes within an existing human performance, but the Suno-generated result appears to reconstruct the character of the voice and the way the song is delivered. The original recording becomes raw material for an entirely new performance rather than simply being brought into tune.

Whether the video was intended as a joke, a defence of his creative process, or a middle finger to the musicians who have criticised him, it lands less like a clapback and more like a demonstration of their argument.

Fawaz has faced mounting criticism over his cover of Madonna’s “Like a Prayer”, which spent four weeks at the top of Australia’s national radio airplay chart and had already amassed more than 38 million Spotify streams by last week.

Following an ABC investigation into the song, its Spotify credits were updated to acknowledge the use of AI for both its vocals and drums.

Fawaz had previously been less forthcoming about the exact role AI played in his work.

Per News.com.au, Australian producer Mitch Thomas, who performs under the name Needs No Sleep, has been questioning Fawaz’s production methods since May. Responding to Thomas on Instagram on June 23rd, Fawaz wrote, “Stop having a little sook buddy, just be better.”

“I use AI as a tool, it’s not that deep,” he continued, adding that he had been releasing music long before AI and was primarily concerned with “providing my listeners with good music”.

Thomas argued that the audio itself contained hallmarks associated with generative platforms such as Suno, including unusual artefacts and a frequency cut off at around 20,000 hertz.

He also compared the track with music by US producer MK and said he would publicly apologise if his assessment was proven wrong.

Fawaz’s new video now offers the clearest public glimpse yet of his workflow.

Using AI as a “tool” is not, by itself, particularly remarkable. Artists and producers increasingly use machine-learning technology for studio tasks, including audio restoration, mixing, mastering and isolating stems.

The more relevant question is what that tool is being asked to do, and how much of the creative lifting it is doing.

While Fawaz may have intended to silence his critics, showing a generative AI platform like Suno turning his self-described terrible singing into a commercially polished vocal is unlikely to settle concerns about the loss of human artistry.