Three Grammy-winning engineers debate whether AI will replace music producers

None of them knew the questions in advance. Each staked out a position early. Brathwaite leaned positive, Diaz sat on the fence, and Reid played sceptic. By the end, those positions had almost entirely flipped.

Why AI mistakes work differently to human ones

The strongest thread ran through imperfection. Reid pointed out that half the records people love contain an error that became the identity of the track. Someone fumbled a part on the MPC and it stuck.

AI does not produce errors that way. When a model gets something wrong, nobody calls it a happy accident. They call it a hallucination.

Diaz pushed the point further. You can prompt a mistake deliberately, but the organic ones come from instinct rather than instruction. All three landed on the same conclusion. As AI output floods the market, demand for demonstrably human work should climb rather than fall.

Diaz added that he and Brathwaite already identify AI-generated files from clients within seconds, purely from the artifacts.

How AI rescued an unusable film scene

Reid’s most concrete example came from outside music entirely.

An independent film company asked him to check their audio. The lav mic on the lead actress had failed during a two-hander, leaving more air conditioning noise than dialogue. Conventional noise reduction could not touch it.

So he pulled her clean vocal from a different scene and cloned her voice. Then he read the lines himself into his MacBook Pro mic and substituted her cloned delivery for his own. The result was seamless, and the actor never found out. Before AI, that fix meant booking studio time and bringing the performer back for overdubs.

Why cheaper tools historically made engineers busier

Brathwaite has watched this cycle before. When Pro Tools arrived, colleagues asked whether his job was about to disappear. Instead his workload grew.

His reasoning is straightforward. Lowering the barrier increases how much music gets made, and more music means more demand for people who can finish it properly. He cited artists who learned to record themselves at home, arguing their productivity translated directly into work for him.

He drew the same parallel with the 1980s keyboard boom. That produced a wave of one-hit wonders before artists like Prince used the same machines to make great records. Diaz compared it to sampling, which drew accusations of cheating before becoming completely normalised.

Where AI mastering fits, and who it serves

The panel took a notably practical line on automated mastering. Brathwaite argued the market will tier by budget rather than collapse. Artists who can afford a named mastering engineer will keep paying for one, because that person’s taste and process are the product.

Everyone else now has a viable alternative. Diaz said he actively points artists toward online mastering services when they cannot afford the engineers he would normally recommend. He does not regard that as a compromise. All three treated wider access as a net positive, in line with how AI has reshaped electronic music more broadly.

The pace of AI labs changed one engineer’s mind

Diaz began the session as the optimist and finished considerably darker. Three new AI labs launched in the six weeks before filming, and the acceleration unsettled him. Asked what he would do if mixing became irrelevant within a decade, he admitted he had no answer.

Brathwaite offered the counterargument. Commercial aircraft largely fly themselves, yet pilots remain in the seats. Machines calibrate on input and output, while humans act on instinct when no input exists. Solving human problems, he argued, is still a human job.

Watch the full discussion on the Waves Audio YouTube channel.