Tag: artificial intelligence

The next 3 interfaces for music’s near future

Posted on | by Bas Grasmayer

Our changing media reality means everyone in music will have to come to grips with three important new trends.

Understanding the music business means understanding how people access, discover, and continuously listen to music. This used to be the record player, cassette player, radio, cd player, and now increasingly happens on our computers and smartphones. First by playing downloads in media players like Winamp, Musicmatch Jukebox, or iTunes, but now mostly via streaming services like Spotify, Apple Music, but also YouTube.

Whenever the interface for music changes, the rules of the game change. New challenges emerge, new players get to access the space, and those to best leverage the new media reality gain a significant lead over competing services or companies, like Spinnin Records‘ early YouTube success.

What is a media reality?

I was recently talking with Gigi Johnson, the Executive Director of the UCLA Center for Music Innovation, for their podcast, and as we were discussing innovation, I wanted to point out two different types of innovation. There is technological innovation, like invention, but you don’t have to be a scientist or an inventor to be innovative.

When the aforementioned categories of innovations get rolled out, they create new realities. Peer-to-peer technology helped Spotify with the distribution of music early on (one of their lead engineers is Ludvig Strigeus, creator of BitTorrent client utorrent), and for this to work, Spotify needed a media reality in which computers were linked to each other in networks with decent bandwidth (ie. the internet).

So that’s the second type of innovation: leveraging a reality created by the proliferation of a certain technology. Studios didn’t have to invent the television in order to dominate the medium. Facebook didn’t have to invent the world wide web.

A media reality is any reality in which innovation causes a shift to a new type of media. Our media reality is increasingly shifting towards smart assistants like Siri, an ‘internet of things’ (think smart home), and we’re creating, watching, and interacting through more high quality video than ever before.

Any new media reality brings with it new interfaces through which people interact with knowledge, their environment, friends, entertainment, and whatever else might be presented through these interfaces. So let’s look at the new interfaces everyone in music will have to deal with in the coming years.

Chatbots are the new apps

People don’t download as many apps as they used to and it’s getting harder to get people to install an app. According to data by comScore, most smartphone users now download fewer than 1 app per month.

So, in dealing with this new media reality, you go to where the audience is. Apparently that’s no longer in app stores, but on social networks and messaging apps. Some of the latter, and most prominently Facebook Messenger, allow for people to build chatbots, which are basically apps inside the messenger.

Companies like Transferwise, CNN, Adidas, Nike, and many airlines already have their own bots running on Messenger. In music, well-known examples of artist chatbots are those by Katy Perry and DJ Hardwell. Record Bird, a company specialized in surfacing new releases by artists you like, launched their own bot on messenger in 2016.

The challenge with chatbots is that designing for a conversational interface is quite different from designing visual user interfaces. Sometimes people will not understand what’s going on and start requesting things from your bot that you may not have anticipated. Such behaviours need to be anticipated, since people can not see the confines of the interface.

Chatbots are set to improve a lot over time, as developments in machine learning and artificial intelligence will help the systems behind the interfaces to interpret what users may mean and come up with better answers.

VUIs: Alexa, play me music from… uhmm….

I’ve been living with an Amazon Echo for over a month and together with my Philips Hue lamps it has imbedded itself into my life to the extent that I nearly asked Alexa, Amazon‘s voice assistant, to turn off the lights in a hotel room last weekend.

It’s been a pleasure to trade in the frequent returns to touch-based user interfaces for voice user interfaces (VUIs). I thought I’d feel awkward, but it’s great to quickly ask for weather updates, planned activities, the time, changing music, changing the volume, turning the lights on or off or dimming them, setting alarms, etc. without having to grab my phone.

I also thought it would be awkward having friends over and interacting with it, but it turns into a type of play, with friends trying out all kinds of requests I had never even thought of, and finding out about new features I wasn’t aware of.

And there’s the challenge for artists and businesses.

As a user, there is no home screen. There is nothing to guide you. There is only what you remember, what’s top of mind. Which is why VUIs are sometimes referred to as ‘zero UI’.

I have hundreds of playlists on Spotify, but through Alexa I’ve only listened to around a dozen different playlists. When I feel like music that may or may not be contained inside one of my playlists, it’s easier to mentally navigate to an artist that plays music like that, than to remember the playlist. So you request the artist instead.

VUIs will make the branding of playlists crucial. For example, instead of asking for Alexa to play hiphop from Spotify, I requested their RapCaviar playlist, because I felt the former query’s outcome would be too unpredictable. As the music plays, I’m less aware of the artist names, as I don’t even see them anymore and I hardly ever bother asking. For music composed by artificial intelligence, this could be a great opportunity to enter our music listening habits.

The VUI pairs well with the connected home, which is why tech giants like Google, Amazon, and Apple are all using music as the trojan horse to get their home-controlling devices into our living rooms. They’re going to be the operating system for our houses, and that operating system will provide an invisible layer that we interact with through our voice.

Although many of the experiences through VUIs feel a bit limited currently, they’re supposed to get better over time (which is why Amazon calls their Alexa apps ‘skills’). And with AI improving and becoming more widespread, these skills will get better to the point that they can anticipate our intentions before we express them.

As voice-controlled user interfaces enter more of our lives, the question for artists, music companies, and startups is: how do we stand out when there is no visual component? How can you stay top of mind? How will people remember you?

Augmented reality

Google Glass was too early. Augmented reality will be nothing like it.

Instead of issuing awkward voice commands to a kind of head mounted smartphone, the media reality that augmented reality will take shape in is one of conversational interfaces through messaging apps, and voice user interfaces, that are part of connected smart environments, all utilizing powerful artificial intelligence.

You won’t have to issue requests, because you’ll see overlays with suggested actions that you can easily trigger. Voice commands are a last resort, and a sign of AI failing to predict your intent.

So what is music in that reality? In a way, we’re already there. Kids nowadays are not discovering music by watching professional video productions on MTV; they discover music because they see friends dancing to it on Musically or they applied some music-enabled Snapchat-filter. We are making ourselves part of the narrative of the music, we step into it, and forward our version of it into the world. Music is behaving like internet memes, because it’s just so easy to remix now.

One way in which augmented reality is going to change music, is that music will become ‘smart’. It will learn to understand our behaviour, our intentions, and adapt to it, just like other aspects of our lives would. Some of Amazon Alexa‘s most popular skills already include music and sound to augment our experience.

This is in line with the trend that music listeners are increasingly exhibiting a utilitarian orientation towards music; interacting with music not just for the aesthetic, but also its practical value through playlists to study, focus, workout, clean the house, relax and drink coffee, etc.

As it becomes easier to manipulate music, and make ourselves part of the narrative, perhaps the creation of decent sounding music will become easier too. Just have a look at AI-powered music creation and mastering startups such as Jukedeck, Amper, and LANDR. More interestingly, check out Aitokaiku‘s Vimu, which lets you create videos with reactive music (the music reacts to what you film).

Imagine releasing songs in such a way that fans can interact and share them this way, but even better since you’ll be able to use all the data from the smart sensors in the environment.

Imagine being able to bring your song, or your avatar, into a space shared by a group of friends. You can be like Pokemon.

It’s hard to predict what music will look like, but it’s safe to say that the changes music went through since the proliferation of the recording as the default way to listen to music are nothing compared to what’s coming in the years ahead. Music is about to become a whole lot more intelligent.

For more on how interfaces change the way we interact with music, I’ve previously written about how the interface design choices of pirate filesharing services such as Napster influence music streaming services like Spotify to this day.

If you like the concept about media realities and would like to get a better understanding of it, I recommend spending some time to go through Marshall McLuhan‘s work, as well as Timothy Leary‘s perspective on our digital reality in the 90s.

Four of the biggest opportunities for the future of music consumption

Posted on | by Bas Grasmayer

A reflection on key trends in music, tech, and user interfaces.

Soundcloud is saved, for now. On top of whatever strategic decisions they make to be able to attract follow-up investments, they face the difficult task of preserving their user community’s trust and winning back part of the trust they already lost. Tumultuous times are ahead, which will be frustrating, but also very exciting as it creates opportunity for new innovation and startups to claim their piece of the pie.

Underserved early adopter: the Myspace moment

Back in April I wrote about the fact that music is about to experience another Myspace moment. What I mean by that is that when Myspace hit decline, as it lost its community’s trust, new platforms got a chance as early adopters bailed and moved on. Musicians started building up audiences on Facebook and Twitter, and sharing their music on Soundcloud.

Now we see another Myspace moment: Spotify is focusing on mass audiences, and the prime early adopter platform has a distressed community due to the continuous struggles that Soundcloud has faced over the last years.

This creates opportunities for concepts such as:

  • Connecting groups of music listeners based on music taste or curiosity:
    • Soundcloud‘s struggling with this due to its failure to keep its search & tagging feature useful as the amount of content grew over the years, and they killed their groups feature;
    • Spotify has deprioritized user-created playlists and removed messaging functionality.
    • TheWaveVR could be one of the startups to fill this gap.
  • Collaboration and feedback:
    • If people are leaving Soundcloud, they need to take that somewhere else.
    • Audiu, which was one of the hottest startups at Sonar+D this year, could play a big role here.
  • Promo services for people who need an easy way to share music to journalists, labels, etc.

You could come up with a lot more ideas and find startups striving to make a meaningful impact there.

A third device in our midst: the Voice User Interface (VUI)

I’ve recently been playing around with an Amazon Alexa I ordered. At first I was skeptical and thought it would always feel awkward, but you get used to it fast and the convenience of a voice-controlled device in the living room (and other rooms) is bigger than I expected. I thought all those times you have to grab your mobile phone, or look something up on the computer, were minor and infrequent inconveniences. Now, the VUI has embedded itself into my life and all kinds of small habits, patterns and every day rituals.

VUIs are going to be the third device: first came PCs (plus laptops), then came smartphones (plus tablets), and now we’re going to get a third addition through voice-controlled devices like Amazon Alexa, Google Home, Apple‘s Siri-based devices, devices in the car, etc. Perhaps this is why Tesla is in talks to do a music streaming service: music is the way into these spaces.

So what happens to the way we browse and explore music when we take the visual user interface away? What place does the smartphone get? What place does the laptop get? And what behaviour extends to our smart speakers?

What happens in AI is very important for VUI apps, but also for chatbots.

Conversational interfaces: the rise of messaging apps

Messaging has frequently been called the next major platform. It enables chatbots, which are apps that live on conversational platforms (this is a trend that’s also strengthened by VUIs). Some of the biggest social platforms to rise up over the last decade were primarily messaging apps, such as  Snapchat, Whatsapp, Telegram, and Kik.

The next step of the social web is messaging, but smarter than the AIM, ICQ, and chatroom phase of social. Facebook is positioning Messenger in such a way that it can live as a platform on its own.

Read Music Ally‘s write-up of the chatbot panel I moderated at Midem.

Short-form video

I urge people to try out Instagram Stories and figure out what it takes to make good content for it. Short-form video content is so important in an age of short attention spans. Some of the hottest platforms to emerge among teens in the last years have been Snapchat and Musically, both limiting the time-length of videos being shared on the platforms. It’s fun, fast, and requires low commitment: making users share and explore more content.

I firmly believe this is going to change the way we write songs and structure them. We’ve already seen how the streaming playlist economy made tracks shorter, with people moving the vocals to the start of the track in order to make skips less likely. In the next years, the video story format is going to strongly impact music.

Instagram is another platform that may fare very well from the decline of user trust in Soundcloud‘s community.

 

I’ll be discussing more of these trends in my newsletter, which goes out every week on Monday. Sign up to stay in the loop.

AI-created non-human music will need human narratives

Posted on | by Bas Grasmayer

To me, it’s beyond a doubt that we’ll all be listening to AI-created music within a few decades, and probably much sooner. The most important way in for this type of music is mood playlists. After the first couple of songs on such playlists, most people tune the music out and get back to their main activity. Does it really matter who has created the song then? Does it matter whether they’re alive? Does it matter whether they’ve ever been alive at all?

[EDIT Aug 15: a small disclaimer since a piece linking here makes an incorrect claim. I don’t think all AI-created music needs a human narrative. I believe the future contains a lot of adaptive, and generative music. More on my point of view in this piece: Computers won’t have to be creative]

We are all creative, and therefore I think it doesn’t matter whether computers will be able to be creative. We are creative as listeners. Computers will be able to predict what we like, then test thousands of versions on playlists until they have the exact right version of the song. As a matter of fact, AI offers the prospect of personalized music, or music as precision medicine as The Sync Project calls it.

A point that’s made often is that AI-created music lacks part of the story people expect with music. People bring it up as an obstacle that can’t be overcome, but it feels like that’s just because of a decision to stop thinking as soon as the point is brought up. Let’s think further.

For one, I think AI-created music already is and will continue to be born in collaboration with people. People will increasingly take the role of curators of music created through algorithms. Secondly, why not give music a story?

Last week at IDAGIO Tech Talks, the music streaming service for classical music where I’m Product Director, we had the pleasure of hearing Ivan Yamshchikov talk about his neural network capable of music composition. With his colleague, Alexey Tikhonov, they fed their system 600 hours of compositions and had it compose a new work in the style of Scriabin. The human narrative was added at the end: as it was performed live by acclaimed musicians (see below).

This is how you get people to knowingly listen to music by artificial intelligence. Most consumption of AI music will be through ignorance of the source of the music. Yet people will warm up to the idea of AI being involved in the music creation process, just like they warmed up to electric guitars, samplers, and computers being used as instruments.

And that’s the narrative that will make it human: artificial intelligence as an instrument which requires a whole new skill set for artists to successfully work with it, and evoke in listeners what they want to.

My Midem wrap-up: Chatbots + marketing Run The Jewels panels

Posted on | by Bas Grasmayer

What a week. I spent it at Midem – one of the most well-known music business conferences organised every year in Cannes. Before I’m off to SonĂĄr+D this week, I thought I’d type up a little update.

About 10 months ago, Midem‘s conference manager got in touch with me to see if we could put a panel together. We landed on the topic of chatbots and Messenger apps, because I think the trend signifies an important shift to a new generation of user interfaces (especially considering voice-activated UI, which will quickly be permeating our daily lives).

It was so great to finally be able to have all these people in the same room, and talk about what they’re doing, get their thoughts out, get them discussing with each other. And the line-up was awesome.

Panel: Messaging Apps, Bots, AI & Music: A New Frontier of Fan Engagement

A quick look at the line-up:

  • Ricardo Chamberlain, Digital Marketing Manager, Sony Music Entertainment (USA)
    Runs a very interesting label bot, which includes messages from artists such as Enrique Iglesias. He also worked on the CNCO campaign with Landmrk, which I’m a big fan of.
  • Luke Ferrar, Head of Digital, Polydor (UK)
    Launched the first chatbot with Bastille.
  • Gustavo Goldschmidt, CEO & Co-Founder, SuperPlayer (Brazil)
    Runs Brazil’s biggest streaming service which not only recommends music through a chatbot, but also builds chatbots for artists, which then drives fans to their service when they want to stream something.
  • Syd Lawrence, CEO & Co-Founder, The Bot Platform (UK)
    Launched the Hardwell bot, which is probably the most well-known example of chatbots being used in music.
  • Tim Heineke, Founder, POP (Netherlands)
    Used to run a cool startup named Shuffler.fm which turned blogs into radio stations and became a kind of StumbleUpon for music discovery, and also co-founded FUGA.
  • Nikoo Sadr, Interactive Marketing Manager, The Orchard (UK)
    One of the most brilliant minds in digital marketing, in general. Previously with Music Ally.

FULL VIDEO:

WRITE UP:

Messaging, bots, and AI’s music evolution by Music Ally’s Eamon Forde

Run The Jewels’ Marketing Panel

A few weeks ago, I was asked if I could also moderated the RTJ marketing panel — which would have been a no-brainer anyway, but having a personal connection to this, made me so excited to do it that I forgot to even introduce myself during the panel.

My first real music business job was with a startup called official.fm. As a student, I listened to a lot of underground and indie hiphop, which made me a big fan of the Definitive Jux label, which put out music by Aesop Rock, Mr. Lif, RJD2, and El-P (also one of the founders). The other founder was Amaechi Uzoigwe, who now manages Run The Jewels. I remember feeling a little starstruck at the time. Now, years later, it was so good to catch up with Amaechi and the inspiring success he’s created for RTJ and himself.

Also on the panel was Zena White, who’s MD of The Other Hand, and does great things for RTJ, Stones Throw, Ghostly, BadBadNotGood, DJ Shadow and more.

FULL VIDEO:

WRITE UP:

How Run The Jewels found fame & fortune: by focusing on fans by Music Ally’s Stuart Dredge

Computers won’t have to be creative

Posted on | by Bas Grasmayer

Every discussion about creative AI sooner or later invites the same objection: “computers will never be able to be creative.”

It’s interesting to think about. One has to define creativity. One has to understand the implications of machine learning and artificial intelligence. Then you have to carefully construct an opinion, and prediction, on whether we’ll ever get to the point of computers being creative – without human involvement.

The problem is, you can think about it for weeks, but at its root it may be as philosophical a question to answer as “does free will truly exist?”

So whenever someone raises that objection I wave it away, because it’s not important whether computers can be creative.

Everyone will agree that human beings have the ability to be creative. We have our imagination. We use it to create: outside us, but also inside our heads. Our perception of the world is creative. It’s why some paintings are enjoyable to us. It’s why children can play with inanimate objects and imagine vast worlds in front of them.

That’s why I believe, no matter what happens, that computers are going to be making great music that can compete with human-created music. Imagine a type of computer-generated Soundcloud, where everything that’s not listened is instantly weeded out. Algorithms will be able to determine what sucks and what will never work, so 90% of the output can be filtered beforehand, or doesn’t even have to be generated.

But for you, as a listener, it’s actually not important whether a computer made the music or a human being… especially when you won’t be able to hear the difference.

Sure, putting on an album by your favourite singer is hard to replace… but there are also moments where you tune into a playlist for sports, for focus, or whatever. Five or ten years from now, can you be 100% sure that all the songs you hear in that playlist are made by humans? Will you think: “Oh, this song is an AI song, because it doesn’t sound creative”?

Of course not.

It’s not important whether AI can be creative, when the recipient is creative.

Is killing privacy the best we can do against secondary ticketing?

Posted on | by Bas Grasmayer

In its push to become a data-driven business, event organisers smell opportunity by connecting ticketing to real identities.

It’s estimated that the market for secondary ticketing is worth $1bn in the UK alone. It’s a problem for fans and artists, since tickets are often bought in bulk by resellers and sold at a much higher rate to fans. None of that added margin goes to the artists (although there are some allegations…).

Recently, Iron Maiden opted to go ‘paperless’ for their UK arena tour in order to curb ticket touting. With success:

“In 2010, 6,294 tickets appeared overnight on three of the major resale platforms — Viagogo, Seatwave and Get Me In! — on the day of sale. In 2016 this had dropped to 207, all on Viagogo, as Live Nation/Ticketmaster had agreed delist the tour at Iron Maiden’s request.”

The tour didn’t go fully paperless, and paper tickets were available, but came with strict requirements towards the fans:

  1. Tickets must carry the name of the purchaser;
  2. Ticketholder must present ID and credit card at the door.

While effective, this is worrying and certainly not a “victory for concertgoers” as Iron Maiden manager Rob Smallwood called it.

It’s not just ticketing: privacy is under attack from all fronts. Many events have decided to go ‘cashless’, requiring people to top up chips in special event wristbands. This way, you know exactly who is ordering what, where, how much, and at what time of the night. If you’re a large organisation like Live Nation, you can build up an extensive profile of users over time.

Valuable data, which may help secure sponsors for alcoholic beverages and helps you to target fans with specific offers, but that data comes with a great responsibility.

Privacy in the age of artificial intelligence

The first multi-day conference and festival I attended that was nearly completely cashless was Eurosonic Noorderslag, earlier this year. It’s a music business conference and showcase event, and has lots of bands playing every night in nearly every bar and club in its host city, Groningen, in The Netherlands. It presented cashless payments as a convenience (ie. to reduce queues at bars).

I immediately researched ways to opt-out and found no good way. It was possible to ‘anonimize’ your chip, but you still have to charge it with your bank card, which ties your identity to it through the transaction records. I had good reason to opt-out and so do you.

On its own, “Bas entered venue X at 21:03 and drank a beer at bar Y at 21:24” seems like useless information. And it probably is. I’m not from a country or culture that frowns upon alcohol, so I’m unlikely to be blackmailed with such a bit of information. However, it is possible for someone to claim they met me there and try to pull some sort of scam. Or worse, for someone to claim they are me by using anecdotal evidence based on these random bits of data, and then scamming someone else.

Criminals are moving from the higher risk ‘traditional crime’ into ‘cybercrime’ which is perceived as lower risk.

More than how someone might use a specific data point, what we should really be worried about is larger data leaks. There are parties that try to collect all information from big leaks. Some use it for good, like Have I Been Pwned, where you can search your email address to see if your login info of any site has leaked. But some people store it for more malicious purposes.

Over time, patterns can emerge in these data sets. These become easier to identify through machine learning algorithms, which can go through large datasets faster than a person could, and can get better over time at making sense of data. Many great ones are open source, like Google’s TensorFlow.

Now, your attendance of live events and what exactly you do there can be tied to your hacked LinkedIn or Dropbox account. Whoever holds that data has power over you.

Artificial intelligence could be trained to send hypertargeted scam emails, which use all the data available about you to trick you. This could result in ransomware being installed on your computer, which often means your hard drive is encrypted and locked and the key to decrypt your data is only turned over after paying a certain fee (usually done through Bitcoin, which makes it harder to track the perpetrators).

This could happen to your phone, but also to your car, or any other devices which are likely to be connected to the internet in a few years from now.

The important take-away is that the more data someone has about you, the wider their ‘attack vector’ becomes. This means they have more paths to target you. Any data point on its own usually doesn’t have much value, but it’s when large amounts of data get combined that value emerges. Facebook, a data company, has a market cap of nearly $400bn.

Privacy is security

Privacy in music should not be an afterthought

We have learned a lot from events. We’ve learned not to use biker gangs for security. We’ve learned to have first aid staff at festivals that are trained to dealing with the effects of alcohol poisoning and mishaps with drugs. We have come a long way to providing experiences that are exciting and safe at the same time.

Now it’s time to worry about our guests’ safety before they arrive, and after they leave our events. Let me be clear:

  • If you request your guests to sacrifice their privacy for ‘convenience’, and you get hacked, leading to people getting blackmailed or scammed, it is YOUR responsibility;
  • If you request this data from guests, make it clear and easy for them to find out how you’re storing the data, what you’re using it for, and when it will be deleted. Don’t just refer to some boilerplate privacy policy full of legalese;
  • When things go wrong, be honest about it and communicate it immediately, so people can take security measures;
  • Never store data about people for longer than you need it. Not storing data is the best way to prevent it from being leaked.

(small sidenote: if anyone ever sent you a picture or scan of their passport, go delete that file and email now)

What can you do as a fan?

Do whatever best protects your privacy. If it feels like you’re being a pain in the ass by requesting an anonimized wristband, great. You should be a pain in the ass. Pain is a great motivator for change. So by all means, request information about how your data is stored and protected, how long it’s stored, for what purpose, etc.

Perhaps the hardest part is willing to skip concerts that don’t have privacy-friendly options. As a consumer we should understand that solving ticket touting by sacrificing guests’ privacy is not a solution. It just shifts the issue and places an additional cost on the consumer on top of the ticket price.

Event organisers need to find a way to mitigate or at the very least minimize that additional cost. This means ticketing organisations have to take measures to invest in technology which helps protect and secure guests’ privacy. But they need to feel pressure, or pain, in order to that.

Data, for ticketing companies, is the same as it is for malicious hackers: the more data you can get on a person, the more valuable it becomes.