At the start of June, Aux App was awarded a £50,000 grant courtesy of Innovate UK’s Young Innovators Next Steps Award.
Founded by Ben Bowler, Aux App is, of course, the music collaboration and file sharing platform for producers, musicians and podcasters, which was launched in beta last year. It provides artists with tools to backup and sync music files and collaborate with other artists, as well as providing access to music industry contacts.
The new grant money has already been earmarked for the development of AI technology that will run throughout the platform. According to the company, tracks and DAW files will be tagged by this AI automatically, allowing deep filtering and searching of a musician’s music projects. Ultimately, this will allow the Aux app to learn the style of a user’s productions and generate ‘connection suggestions’ in the wider community based on that profile.
From playlisting to discovery to data analysis and even artistic creation, artificial intelligence is playing an ever-greater role in the music biz – and it’s something that the majors built into their own systems a long time ago. Bowler’s mission, he says, is to put some of that power into the hands of independents.
Here, we ask Bowler about Aux, his AI vision and his take about the direction the music industry is heading generally…
What problem will this AI solve for artists?
Throughout the Aux app beta program, I’ve been speaking to musicians every single week and it’s been really interesting to see the type of musician the project has attracted early on. Namely, early-stage artists, solo artists pre-first release, bands with their first album and a small live following, even music students sharing their college projects. As I’ve been speaking to these artists about the problems they face, a number of themes have come up again and again. The most common is the need for connections – creative connections in form of collaborators or industry connections to help them take the next step in their music careers. This is why from the very start we began building a guide of music businesses into the app including mastering engineers, music studios, labels and publishing companies. All of these businesses can mean the difference between an emerging artist making an impact on their scene or disappearing into the ether.
"The end goal is a little bit of AI sprinkled across the whole biz. It’s a technology that can be used to improve and enhance any process that includes technology in some form already today."
What problem will it solve for the music industry?
From the industry side there’s the reverse problem. Music files are the core asset of the recorded music industry, yet many music companies I speak to have a cobbled together system for managing the volume of music files they work with day to day. The more public facing, the larger the problem, with tracks coming in from multiple angles: SoundCloud links, MP3s via Whatsapp and email. For many it’s got to the point that they only listen to tracks that come from trusted contacts. This is where I want to change things. I know that the ability to create great music is not correlated with having a mate that happens to work at a record label or publishing company. There is great music out there, if only you could find it without spending days sifting through every single track that comes into your inbox, or what about looking beyond your track inbox, finding artists before they’ve even sent you anything. This is the opportunity for Aux’s AI tech.
What impact do you think AI is having in music currently, more broadly?
The group of AI technologies that are being hyped today (Neural Networks) are super good at solving prediction problems. What I find fascinating is how many problems in music can be reduced to prediction problems. Give the AI access to a million tagged songs to train, and it can tell you what instruments and genre your track is, highly accurately. Give the model millions of user playlists and it can generate a new playlist made for you like Spotify’s Discover Weekly. Major labels have had their own internal AI tools for years, managing their own content, making suggestions for sync projects, tracking accelerating growth in upcoming artists social media for A&R, even predicting chart success. Where it starts getting interesting is in machine creativity. These algorithms use an approach called GANs (Generative Adversarial Networks) which again take in existing music content to train but can then generate entirely new creative output in the style of what it’s been given. This is just starting to get out there in the form of plugins that generate unique sounds every time, and sample packs made entirely by machine.
What do you think the end goal is for AI in the music biz? What does maturation look like?
The end goal is a little bit of AI sprinkled across the whole biz. It’s a technology that can be used to improve and enhance any process that includes technology in some form already today. From production, generating creative samples, basslines, lyrics, to auto track tagging that helps you sort and filter all of the music projects you or your company are working on, to making super relevant connections on sites like Aux. Many people see AI as a scary disruptive force out to take your job, but if we do it right in the music biz, I see AI as the ultimate assistant adding a little intelligence to every step of the creative journey. My personal vision for the technology is getting it out of the exclusive hands of the big players and bringing it to the independent artists and labels. I’ve always been more interested in supporting people succeeding on their own terms.
"Aux allows artists, labels and publishers to make connections before they have even released a single track by analysing the work-in-progress DAW project files."
The AI functionality you’re building in Aux almost fulfils an A&R duty, correct?
This is a part of it, the first step – connecting artists to labels and publishers, building new creative relationships and getting new music out into the world. More broadly though, it’s about enabling new connections across the industry as a whole. Technology across many industries has enabled collaborations that would have never been possible before. Industries like video production, graphic design and software development have been revolutionised by the ability to work remotely in truly global teams. In music, it’s still all about getting into the same studio at the same time to record a track. A true remote collaboration tool like Aux, opens the door for global creative collaboration we have not yet seen in the music industry.
What do you think the long-term future of music A&R looks like?
In terms of A&R, it’s all about spotting talent first. Earlier, I talked about major label tools that watch social media followings to find new emerging artists to sign. That’s using signals from an artist after they release their music and start to build hype. Aux allows artists, labels and publishers to make connections before they have even released a single track. It does this through analysing the work-in-progress DAW project files. This means artists can be discovered earlier than ever before. It’s just down to enterprising labels to recognise this and get involved with Aux ahead of the curve. Personally, I’ve always seen the job of a good label as an incubator – taking early-stage artists and supporting them to their future success. Aux’s AI technology helps make this happen.