top of page

Aubrey Whitfield

Oct 11
11 min read

Songwriter, producer and mix engineer based in London & Sheffield


"Yes, it might be cheaper to use Suno to generate a guitar part, but why not get a good guitarist to create something fresh and new?"

Aubrey Whitfield

Woman smiles at a music mixing desk in a purple-lit studio, surrounded by monitors and audio gear.



Aubrey Whitfield is not against AI. She uses it where it saves her time, and she says so without apologising. That is exactly why I wanted her in this series.


She is a producer and mix engineer in the UK, which puts her on the other side of the glass from most of the people I’ve talked to so far. Artists tell me what AI might do to music. Aubrey can tell me what it is already doing, because the files land on her desk. Over the last six months enough of them have been fully AI-generated that she now has a policy for it.


We don’t agree on everything. She thinks audiences don’t care whether a song was made by a machine, and that this won’t change. I’ve left that in exactly as she said it.



Let's start with you.

Tell me a little about yourself and your work as a producer. What kind of music and which kind of artists do you work with, what does your role usually cover (producing, recording, mixing, co-writing), and what does making records mean to you personally?

I would always call myself a songwriter first, then a producer and mix engineer. But I do all three nearly everyday of the week! I do a lot of writing and producing for Production Music and TV shows and that comprises about 70% of my work. The rest of the time I am mixing records for labels and publishers as well as independent artists. I have an eclectic taste in music so I create and mix a wide variety of genres, but my forte is in both modern chart pop and singer-songwriter music. I’ve been doing this full-time for nearly 10 years now and I still can’t believe that I get to wake up each day to create music and get paid for it.


When did AI first show up in your studio?

Was there a particular tool, plugin, client request or moment that made you realize: this could really change how records get made?

I was mixing a song for an artist back in early 2026 and there was a choir stem that sat behind the lead vocal. It sounded like a huge, lush church choir. It sounded great but I thought the recording was a bit ‘off’ and perhaps they had exported it as a low quality Mp3. When I asked the artist he said that the choir stem had been generated by Suno. I was quite amazed by the fact it could generate a choir stem with that level of realism. So I decided to try out Suno myself and I uploaded a couple of old demos that I never really finished because I didn’t know what to do with them, and I was blown away by the results - the songs sounded great - really excellent instrument choices and top class vocal performances. But the problem for me was the sound quality of the songs. The production was top-level, but the sound quality was well below average for the standard needed for professional records. So I haven’t used it since. But it was this moment that made me realise that AI has come a long way in a short amount of time.


Which AI tools are part of your workflow today, if any?

This could be anything from stem separation, audio repair and mix or mastering assistants to vocal tools, sample or idea generation, and the admin side of running a studio. If you've tried something and stopped using it, or you deliberately keep AI out of the room, I'd be equally interested in why.

I’m quite open to AI tools if they speed up my workflow as I’m a super busy producer! I use Chat GPT to help me identify what’s trending at the moment in terms of music productions and mixes, as well to generate some chord sequences that might work for a particular style of song. I have also used tools such as Illugen from Waves Audio where you type in a prompt and it generates a brand new sound just for you, and same with Co-Producer from Output. Sometimes I will use stem separation tools from the likes of LALAL-AI if a client has sent me some stems with effects that need removing or to separate a kick from a snare for example. I think as a producer who works in the mainstream, it’s imperative to keep on top of trends and evolving technology, so understanding how popular AI generated tools like Suno work is important. Suno is not something I would use personally because I already have the skills in writing, producing and mixing, but I can see why it is appealing to writers who struggle to create productions. That said, in my professional opinion, the quality of AI generated songs has a huge mountain to climb before they reach a professional level.


Where do you personally draw the line?

Which parts of a production are you happy to let a machine help with, and which would you never hand over, even if the result sounded just as good?

For me, I think it’s morally important to work with real musicians and vocalists where possible. Yes it might be cheaper to use Suno to generate a guitar part, but why not get a good guitarist to create something fresh and new? Also, I don’t see anything wrong in perhaps using AI to generate some production ideas if you’re feeling a little stuck and then using those ideas to create a new song in your DAW. I think AI generated music can be very helpful in this sense. I am generally quite supportive and open to new technology and ways of doing things, but I personally would never produce or write a song with AI - even if it’s quicker and might get me good results - I am an experienced producer and I want to keep my productions pushing boundaries, and to be unique and not necessarily based on the commercial success of thousands of other records.


What are you seeing from the artists and clients you work with?

Do people arrive with AI-generated demos, references or ideas? Has it changed what they expect from you in terms of speed, budget or the role you play?

Over the last 6 months I have had a significant number of non-label artists come to me to mix their songs that have clearly been completely AI generated. To me, it’s very obvious which songs have been created with Suno - the vocals have a lossless file sound quality to them. For a mix engineer, it is very difficult to improve AI generated songs in mixing - the stem separation by Suno further degrades the audio of the AI generated stems and they are caked in effects that I can’t remove even using effect removal tools. Mixing rarely improves what is already there with AI generated songs. Only mastering can help enhance the sound a little. So I now have a policy that I don’t mix songs that have been completely AI generated. Part AI generated is ok, but I don’t want creators paying me for something that is not really going to sound any better.


What does "producing" actually mean in the age of generative AI?

If someone enters a prompt and gets back a finished track with arrangement, instruments, vocals and mix, is that a production in the same sense as a record made by people in a room? Why or why not? And what, if anything, does a producer add that a prompt doesn't?

This is the crux of the debate in the music industry right now. My view is that it is not producing. It’s more like executive producing where you are responsible for the high-level creative direction of a project, rather than a hands-on technical role. The AI is the producer in my opinion whereas the user is a creator. But these definitions are loose and will surely change over the coming years. If we look at the definition of producing, however, it actually says: “Music Producing is the full process of turning a raw musical idea, melody, or feeling into a finished, polished song that can be heard and enjoyed by an audience”. If we look at this definition you could argue that an AI creator does fall into the category of a music producer. I’m sure we all have different views on this and will continue to as AI evolves!


What excites you about AI, and what worries you?

Where could it genuinely make you better or faster at your job? And on the other side, which developments concern you most?

I think there’s a lot about AI that is positive. I can use it to speed up my workflow, assist my musical ideas and help me create a better production all round. But the insanely rapid development of AI is concerning to me. Not just in terms of AI becoming popular for music creation affecting the incomes of producers and engineers, but more that the music landscape will lack heart and soul. If more producers are turning to AI generation tools like Suno, the charts could become saturated with AI generated music. And AI generated music has a recognisable quality about it. I can instantly tell if it’s AI generated. If we don't rely on musicians, writers and producers to revolutionise music trends and develop new trends that push boundaries like we have done for decades, then we’re going to have music that sounds similar and is stagnant - with a huge slow down of developing new music sounds and trends.


What do you think AI will do to the business of making records?

Think about studios, producers, mixing and mastering engineers, session musicians, songwriters, labels and independent artists. Who could benefit, and who might have the most to lose? Have you already felt any of it in your own bookings or budgets?

For me, I don’t feel I have lost any work due to AI…yet! People still need high quality productions and mixes but I am seeing an increase in people wanting me to take their AI generated songs and re-create and produce it into something that sounds less AI generated. And that’s where I think we are at the moment - there are people using Suno for fun, and there are serious writers or producers using Suno to generate ideas that then need the intervention of a producer or mix engineer. AI is going to get better. If we think about what it has done in the last 12 months or so, what will it be like in 3 years time? The sound quality issues I’ve talked about will probably have been resolved, the editing functions will be way more capable than they are now. And I find that scary. If someone can generate an incredible sounding production and mix by spending £20 a month by generating a prompt, they might very well choose that option over paying £500+ for a producer, and then a mix engineer and then a mastering engineer. But at the moment there is a real stigma in creating AI generated music and until the sound quality of AI matches what a top-level producer can do, I don’t think AI is going to overtake the much needed skills of a human producer and mix engineer. But as for the future, I am worried.


Copyright, consent and credit: where do you stand?

Should AI companies be allowed to train their models on existing recordings, or should the people who made them have to opt in? If a recording contributes to an AI model, should producers and engineers be credited or compensated alongside the artists and songwriters? And should the use of AI in a production be disclosed in the credits?

I 100% categorically think it’s wrong for AI companies to train their models on existing recordings without a) the permission of the copyright holder and b) compensating the copyright holder for doing so. I think it gets into grey territory when it comes to original producers or writers getting credited on AI generated songs that have been influenced by their work. A lot of traditional producers are influenced by other songs and create songs influenced by songs they find inspirational, and as long as you aren’t copying that original work, the original writers aren’t credited. So I think it’s a fine line. I do think that all or part AI generated songs should be disclosed in the credits. In some ways you could argue does it really matter if it’s AI generated or not if the listener enjoys the song? But I think it shows respect to non AI music producers by showcasing that their work isn’t AI generated - otherwise you might be accused of creating AI generated material when you haven’t (this has happened to me before!)


Imagine someone is 16 years old today and dreams of becoming a producer.

What would you tell them to learn? Is it still worth learning an instrument, training your ears and spending years in studios, or has AI changed the advice you would give them?

In my opinion there is no substitute for being both a musician and a producer. Being able to play an instrument (piano and guitar for me) gave me the tools to understand things like chord progressions, time signatures, key signatures, and how music is constructed. I believe this makes me a better producer. So my advice would be to learn the piano/keys at least. I think we’re a long way off from AI being the leading tool that music creators use for producing and mixing. So my advice would be this: learn how to play keys, learn some basic music theory, learn how to use a single DAW inside out, keep on top of music trends by seeing what’s trending on Shazam, and keep on top of music tech developments by reading music-tech magazines and blogs.


Fast-forward five years.

What do you think a recording session will look like? Will artists still come to a producer and a studio? Will audiences care whether a record was made by humans or AI, and do you think we'll even be able to tell the difference?

I don't think audiences care if a song was made by AI or not. I think it’s producers and musicians who care more about it. And I can’t see that changing if I’m being honest. There will always be people who want to use Suno because they don’t have the musical or producing skills needed, and there will always be musicians and producers that do have the skills or want to work with someone who has the skills. There’s going to continue to be a split for sometime.


The Human Frequency.

What is the one thing a human producer brings to a record that you believe AI can never replace?

AI isn’t very good at doing something new and fresh and groundbreaking. For example, Suno is trained on thousands of hit records and then produces songs based on those hit records. Where is the innovation? The drive to do something different that breaks away from everyone else? Would it ever create a Bohemian Rhapsody? AI doesn’t generate fresh new trends and that’s the one thing that a human producer has the ability to do. Music trends change all the time, yet AI will just replicate what’s popular in those genres, a human producer will continue to push trends and create something fresh. Once AI starts to innovate…. that will be a different story!

Aubrey Whitfield is a British Music Producer, Songwriter and Mix Engineer based in London & Sheffield, UK. She has over 20 years experience as a producer and songwriter; having worked with a diverse array of artists including The Shires, Meek, J-Fla, Nahreel, George Shelley (Union J) and Gaston Dalmau. Songs she has written and produced have entered the Top 40 charts in over 12 countries, generating over 131 million streams worldwide. In 2026 & 2025 she won an SME News Business Award for Mixing & Producing and in 2020 was included on the Alternative 100 Power List.

Woman in a music studio sits in a brown chair beside glowing monitors and audio gear under purple lighting.


Nice to know:

In April 2026, she launched an advanced online mix analysis platform called AubioMix. It has been built to analyse mixes exactly like she does, and provide extensive feedback across 15+ different areas of a mix in minutes, combining her 20+ years of mixing experience with her language and tone. To date, over 10k mixes have been analysed through AubioMix with countless producers improving their mixes and reaching a professional level. Users can upload a mix for free at https://www.aubiomix.com


A few productions or artists she has worked with:

The Shires (Mastering Undone Vol 1 album), Meek, J-Fla, Nahreel and George Shelley (Union J). Her songs have been placed in many TV shows around the world including Traitors, Love is Blind, Love Island, Nike adverts, Film 4, World Cup coverage, Olympics coverage and more.


Upcoming projects and releases:

About to release my first ever solo EP on a major publisher. ‘Fire In The Belly’ is getting released on BMG in the coming weeks.




Social links

 
 
 

Comments


bottom of page