AI in the cultural industries: uses & limitations

Can Artificial Intelligence create music you want to listen to, TV you want to watch, books you want to read, or news you want to consume? If the industry news feeds I’ve been tracking daily are any indication, the answer is increasingly ‘yes’. Although after doing my own research today, I discovered that AI is both more useful than I first assumed and less powerful than media reports often claim.

In today’s edition of Cybercultural, I look across the cultural sectors to examine the real-world impact of AI software.

AI-created music

One thing’s for sure: while AI is potentially a useful tool for the arts, it’s also more prone to hyperbole than the other technologies I cover here. The worst example of AI hype I came across recently was a quote from Silicon Valley billionaire, and noted AI enthusiast, Vinod Khosla. As Techcrunch reported:

“I actually think 10 years from now, you won’t be listening to music,” is a thing venture capitalist Vinod Khosla said onstage today during a fireside chat at Creative Destruction Lab’s second annual Super Session event.

Instead, he believes we’ll be listening to custom song equivalents that are automatically designed specifically for each individual, and tailored to their brain, their listening preferences and their particular needs.

According to Techcrunch, Khosla added that “AI-created music is already making big strides.”

These comments met with howls of derision on Twitter (including from me). Khosla claims he wasn’t fully quoted by Techcrunch, so in the interests of fair play here is what he replied to one of his critics on Twitter:

“Idiots like you take half of what I said. I specifically said we'd have music personalized to each person”

Let’s get real here. AI-created music will never usurp human-created music. If that’s what Khosla was implying, he’s just plain wrong. To many of us, music is about so much more than mere sound in our ears - it’s also about a (human) connection with our favourite artists, the meaning and inspiration we glean from their art, and community with fellow fans.

That all said, AI is increasingly being used as a tool by musicians; and it’s true that some AI-created music has made its way to our streaming music playlists.

Journalist Stuart Dredge wrote an excellent overview of AI in the music industry in February. He pointed out there are a range of use cases for AI, including creating backing tracks for video, helping an artist come up with melodies or lyrics, and automatically creating mood music. He listed some of the technology companies building these types of AI solutions for music:

“Startups include JukedeckAmper MusicAivaWaveAIMelodriveAmadeus CodeHumtapHumOnAI MusicMubertEndel, and Boomy, while teams from GoogleSonyIBM, and Facebook are also looking at what A.I. music can do now and what it could do in the future.”

In an article in The Guardian, Hazel Davis profiled another startup building an AI tool for musicians. A company called MXX “has created what it says is the world’s first AI tech that allows individual users to instantly edit music to fit their own video footage, complete with rises and fades.”

Ken Lythgoe, head of business development at MXX, positioned his company’s technology as a tool to empower artists:

“There are two types of AI – the AI that is here to replace us and AI that is here to empower us – we are definitely in the empowerment camp. We are not about computers replacing musicians or editors; we are firm believers in the creative process.”

MXX is being targeted at advertising agencies, game creators, production companies, music supervisors, labels and sync agencies.

AI in Hollywood

This week in Broadcasting & Cable, two executives from global design and innovation consultancy Fjord argued that “TV producers need to get with artificial intelligence, or get left behind.”

According to Fjord’s Siva Natajaran and Jeff Bauer:

“AI is becoming more mainstream across the entire TV and film ecosystem. From the front stage (such as recommendation engines) through the creative process, scripting, shooting, post-production, and all the way to the back stage (such as meta-tagging and distribution), AI is enabling industry newcomers to leverage new business opportunities.”

The pair acknowledged that AI tools are currently “at too early a stage for studio executives to fully embrace yet,” but they list some “quick efficiency wins AI can already deliver during the creative process.” One example was an automated script-scanning tool called Scriptbook.io, which can be used for creative purposes in addition to more mundane uses like checking for IP infringements:

“Simply by uploading the PDF of a script, Scriptbook.io can provide a detailed analysis on characters, detecting the protagonists and antagonists etc., within minutes.”

It’s a fascinating article, because Natajaran and Bauer convincingly argue that AI can be used for a wide range of use cases in the tv and movie industry: from the early scripting stage, to filming and then production, to distribution, to marketing, and more.

Automated journalism

Along with music, news journalism is where AI is making the most immediate impact in the cultural sectors. There’s even a term for it: “automated journalism.” It’s also known as algorithmic journalism or robot journalism, according to a Wikipedia article.

“Roughly a third of the content published by Bloomberg News uses some form of automated technology,” reported The New York Times in February. Bloomberg uses a system called Cyborg, which is particularly useful for covering company earnings:

“The program can dissect a financial report the moment it appears and spit out an immediate news story that includes the most pertinent facts and figures.”

NYT also listed The A.P., The Guardian and Forbes as other media outlets that use robot reporters to one degree or another.

Lisa Gibbs, the director of news partnerships for The A.P., told The Times that AI isn’t about replacing human journalists. Rather, it’s about taking away some of the more mundane tasks of news reporting:

“The work of journalism is creative, it’s about curiosity, it’s about storytelling, it’s about digging and holding governments accountable, it’s critical thinking, it’s judgment — and that is where we want our journalists spending their energy.”

In the end, as with music, I’d prefer to spend my time reading journalism written by human beings. I can see the benefits of AI - and indeed any automation tool - for data-gathering and data analysis. AI is also clearly useful for ‘just the facts ma’am’ style reporting, such as with company earnings.

I’m obviously biased, but the real value of journalism to me is the added value on top of the facts - the analysis, the context, a worldview, human empathy, and more. All things an AI cannot deliver presently (and probably never will, since AI is fundamentally a different type of ‘intelligence’ than human intelligence).

Infinite monkeys writing books

So if robots can write news articles, can they produce books too? Some publishers are experimenting with this.

In April, The Verge reported that academic publisher Springer Nature had released “what it claims is the first research book generated using machine learning.”

The book is clearly labeled as not being authored by a human writer. It’s entitled ‘Lithium-Ion Batteries: A Machine-Generated Summary of Current Research’. Sounds like a real page turner!

As explained by The Verge’s James Vincent:

“…as the name suggests, it’s a summary of peer-reviewed papers published on the topic in question. It includes quotations, hyperlinks to the work cited, and automatically generated references contents.”

If this reminds you of the infinite monkey theorem (stating that if an infinite number of monkeys were left to bang on an infinite number of typewriters, sooner or later they would accidentally reproduce something William Shakespeare wrote), then you should know that AI-generated academic texts are mostly a gimmick right now. In his review of the batteries book, Vincent pointed out that it included “garbled and incoherent sentences” and was generally difficult to read.

Takeaways

  • AI-created music and ‘automated journalism’ are two emerging trends, and both will fill a need where the creative requirements are low. For instance, maybe a 30-minute AI-generated composition entitled ‘Motivational Mood Music No. 61’ will be better background music while you work than Bob Dylan’s ‘Highway 61 Revisited.’

  • There’s no chance music or articles composed entirely by machines will ever replace the added value delivered by human composers and writers. Vinod the VC can bop to the robots if he likes, but I’ll stick with flesh and blood artists thank you.

  • AI tools are already augmenting workers across the cultural industries. As evidenced by the Hollywood article above, there are AI tools available to assist the creative or business process at all stages of making a tv show or movie. That’s a trend in music, journalism, book publishing, and other cultural sectors too. My bet is that this will be the real value of AI for the foreseeable future - as a way to augment human creativity in the arts.

Lead image: Z-machines; Initial design and prototype for a robot band, by Yuri Suzuki.


Tracking 👀

  • Hypebot contributor Kyle Bylin writes that Spotify's recent acquisition of Anchor suggests it will empower artists and influencers to create their own shows. 🎧

  • This July, Google Chrome will make it easier to bypass paywalls; to which the news media industry replies sarcastically, thanks again Google! 🗞️

  • Labels, streaming services and others announce their support of a new “Code of Best Practices” aimed at preventing and detecting stream manipulation. 🎹

  • Ten Business Models for Indie Authors; good tips for self-publishing authors. 📚

  • With MTV’s growth as guide, Complex targets international expansion. 📹


Data Points 📊

  • Merlin paid indie labels and distributors $845m in the past year, showing strong growth. 💰

  • New MIDiA report suggests that classical music will be streaming’s next genre. 🎻

  • Pew survey of US adults: 66% say social networks should police offensive content but 69% have little to no confidence they will correctly identify such content. 📱


Cybercultural is a new email newsletter that covers the intersection of technology and the cultural industries. It’s currently in beta; paid subscriptions start 1 July.

Your ‘likes’ and shares help get the word out about Cybercultural. Please share by email and/or social media if you enjoyed this article.

Your feedback is much appreciated too. If you’ve received this by email, just hit the reply button. Otherwise ping me on Twitter (@ricmac).