Three Grammy-winning engineers put AI’s biggest promise to the test — and reached a more complicated conclusion about the future of music production.
Artificial intelligence is moving deeper into music production at a pace that would have seemed impossible only a few years ago. From stem separation and restoration to voice cloning, mastering and full-track generation, tools once reserved for expensive studios are becoming accessible to almost anyone with a computer.
But will that accessibility eventually make producers, engineers and mixers obsolete?
That question was at the centre of a recent discussion hosted by Waves Audio, which brought together three highly decorated figures from the recording industry: Leslie Brathwaite, Lu Diaz and Preston “Prizzie” Reid.
Their initial positions could hardly have been more different. Brathwaite approached the technology with optimism, Diaz remained undecided, while Reid entered the conversation as the sceptic.
By the end, however, all three had reconsidered their starting points.
And rather than predicting the death of the human producer, the conversation pointed toward something more nuanced: AI may make production easier, but human instinct could become more valuable precisely because machines are everywhere.
The Value of the “Mistake”
One of the most compelling arguments centred on something machines struggle to understand: the creative value of imperfection.
Reid highlighted a familiar phenomenon in music history. Some of the most memorable moments on beloved records were never planned. A slightly late hit, an imperfect MPC performance or an accidental sound can become part of a recording’s identity because someone recognised its character and decided to keep it.
That is fundamentally different from an AI error.
When an artificial intelligence system produces an unexpected result, it is generally treated as a malfunction or hallucination. When a human musician makes an unexpected move, that same mistake can become inspiration.
The distinction becomes increasingly important as AI-generated music fills digital platforms.
If technically flawless music becomes virtually unlimited, listeners may begin placing greater value on recordings that carry obvious evidence of human decision-making.
Diaz noted that mistakes can technically be requested from AI, but deliberately instructing a machine to make something imperfect is not quite the same as stumbling upon an unexpected idea through instinct.
That difference could become one of the defining characteristics of human-made music in the AI era.
AI Can Fix Problems Humans Couldn’t
Despite the concerns surrounding AI, the engineers were far from dismissing its usefulness.
Reid offered one of the clearest examples.
An independent film production had captured dialogue from a two-person scene using a lavalier microphone. The recording was severely compromised by air-conditioning noise, leaving conventional restoration tools unable to recover the dialogue adequately.
Instead of abandoning the recording, Reid used AI voice technology to reconstruct the performance.
He extracted the actress’s clean voice from another scene, recorded himself delivering the dialogue into his MacBook Pro microphone, and then used the cloned voice to replace his own performance.
The result was reportedly seamless.
Before this technology existed, solving the problem would have required bringing the actor back into a studio for additional recording.
For engineers, this represents the less sensational but potentially more important side of AI: not replacing creativity, but making previously impossible repairs possible.
Technology Has Threatened Music Jobs Before
Brathwaite sees today’s AI debate as part of a much longer technological cycle.
When Pro Tools transformed recording workflows, many engineers feared that digital production would eliminate their jobs. Instead, the opposite happened. As recording became easier, more artists began creating music, which ultimately generated more work for professionals capable of taking those ideas and turning them into finished records.
He sees a similar pattern in earlier technological shifts.
The widespread arrival of affordable keyboards during the 1980s made music production more accessible and contributed to a wave of inexpensive productions and one-hit records. Yet the same technology eventually became a fundamental creative instrument for artists such as Prince.
Sampling followed a comparable trajectory. Initially criticised as a shortcut or even a form of cheating, it eventually became one of the defining techniques of modern music production.
AI could follow the same path.
The technology may lower the barrier to entry, but lowering that barrier also means more people making music — and potentially more demand for specialists who know what to do with it.
AI Mastering Won’t Necessarily Kill the Studio
Automated mastering presents perhaps the clearest example of where the market could split rather than disappear.
AI-powered mastering platforms can now deliver surprisingly capable results at a fraction of the cost of a traditional mastering session.
Brathwaite believes that will create different levels of service rather than eliminate professional mastering altogether.
Artists operating on limited budgets can access automated tools, while established artists and labels can continue paying for respected mastering engineers whose value extends beyond technical processing.
In that environment, the engineer’s taste, judgement and reputation become part of the product.
Diaz takes an equally practical approach. He already recommends online mastering solutions to artists who cannot afford the engineers he would ordinarily suggest.
For him, accessibility is not necessarily a threat.
It can be an entry point.
The Speed of AI Is What Changed the Conversation
Perhaps the biggest concern came from Diaz, who entered the discussion relatively optimistic about AI but became significantly more cautious during the conversation.
His concern was not simply what AI can do today, but how quickly its capabilities are improving.
The emergence of several new AI laboratories in the weeks preceding the discussion made the pace of development particularly difficult to ignore.
When asked what he would do if mixing were to become largely irrelevant within the next decade, Diaz admitted that he did not have a clear answer.
That uncertainty may be more revealing than any prediction about AI replacing producers.
The technology is evolving quickly enough that even professionals working at the highest level cannot confidently map out where the industry will be ten years from now.
Why Human Instinct May Become the Ultimate Advantage
Brathwaite offered perhaps the most optimistic counterpoint.
He compared AI’s development to modern aviation. Commercial aircraft can automate enormous portions of the flying process, yet pilots remain essential.
The reason is simple: automation excels when the parameters are known.
Human beings become most valuable when something unexpected happens.
That principle translates directly into music.
AI can analyse patterns, generate variations, imitate styles and optimise technical processes. But production is not always about choosing the mathematically best result.
Sometimes the best decision is the strange one.
The mistake.
The emotional take.
The sound that technically should not work but somehow does.
And that may be the real dividing line between automation and artistry.
AI Could Make Human Music More Valuable
The discussion ultimately avoided both extremes of the AI debate.
It did not suggest that artificial intelligence is harmless, nor did it predict an inevitable future in which human producers disappear.
Instead, it highlighted a potentially more interesting possibility.
As AI makes technically competent music cheaper and easier to produce, human creativity could become the premium layer.
The producer who understands when to break the rules may become more valuable than the tool that knows every rule.
For electronic music, where technology has always been closely tied to creativity, that distinction could become particularly significant.
AI may change how records are made.
But the question of why a particular sound should exist in the first place remains much harder to automate.