It’s revolution time again in the entertainment business.
[chanting]
Big Data is bringing big change and faster than ever before.
AI has the potential to disrupt every corner of the entertainment industry.
A new wave of artificial intelligence is beginning to transform the way entire industries operate.
We can build an algorithm that listens to all the music in the world.
Determine what we listen to and watch.
The AI predicts how much money this film will make.
And generate new forms of art.
Generative AI moves from being something that I think’s quite beautiful to occasionally a little bit intimidating.
So who are the winners?
AI driven names were two of the top stocks in the S&P 500.
And losers.
Generative AI is an existential threat to the TV and film industry.
This is a real earthquake, not just for the entertainment industry but for our lives in the 21st century.
NOW&NEXT
How AI is generating a revolution in entertainment
[singing] I’m losing hope, I tell you baby, I’m going under, I can’t help…
In this studio an artist, Alfie, is playing a new song to the boss of his record label, Ankit.
Alfie Castley
Ankit, could you hit spacebar just so I can see my levels please.
It’s the kind of scene that’s been typical of the music business for decades.
Alfie Castley
We’re good to go, I think.
[singing]
But it was not Ankit who discovered Alfie, it was AI.
Ankit Desai, Founder and CEO
Snafu Records
Alfie Castley was one of the artists that we found through our algorithm.
Ankit’s label is based in Sweden and uses AI to find new artists across the world. AI hit the right note when it identified that London-based Alfie had a big fanbase thousands of miles away on another continent.
Ankit Desai, Founder and CEO
Snafu Records
He is an artist from the UK. His five biggest cities in terms of where his fans come from, are all in Southeast Asia.
[singing]
Alfie’s biggest hit, ‘Teenage Mona Lisa’ had about 700,000 streams in his home country but 7 million in Indonesia and over 16 million in the Philippines in its first year.
Ankit Desai, Founder and CEO
Snafu Records
It’s been viral here in the Philippines for good lord knows how long.
Technology helps you kind of like find these let’s say counterintuitive patterns.
I think this is the one that, that both Carl and Savan and Ramy…
Ankit’s company is using AI to get better at spotting opportunities and trends in the digital era. AI can analyse metadata from social media and the behaviour of viewers and listeners faster than other technologies.
Ankit Desai, Founder and CEO
Snafu Records
We look across Spotify, YouTube, SoundCloud, any platform that has data. You’re looking for signs of someone raising their hand and saying, “I think this is a good song.” It could be in addition to a playlist, a comment on TikTok, the footprints in the sand. The biggest use that we’ve had for AI has been how do we take this just fire hose of content and distil it down to something that’s manageable on a human level.
Alfie Castley
Yeah, I think let’s call it, if that’s alright guys.
Back in the day, the whims of artists and music executives dictated what songs the public would hear. Digital technology, boosted by AI, has now turned that model on its head.
Ankit Desai, Founder and CEO
Snafu Records
You’re turning hundreds of thousands, sometimes millions of people into talent scouts. You’re democratising the selection process to the fans.
Alfie Castley
You like?
Ankit Desai, Founder and CEO
Snafu Records
Oh my god. That was amazing.
Alfie Castley
Oh thank god.
AI is the latest chapter in the story of how Big Data has completely transformed the music business over the past decade. The arrival of streaming services changed everything from how profits are made to how songs are structured.
Tom Wainwright, Media Editor
The Economist
Musicians now get paid per play and so there’s a particular incentive to make sure that songs sound good right at the beginning so that people don’t skip them.
Big Data mining also completely disrupted the world of video. Netflix, which launched its streaming service in 2007 led the way in using data to choose what content to make and who to serve it up to.
Neta Alexander, Assistant Professor of Film and Media Studies
Colgate University
On Netflix, everything you do in the platform is a data point. Whether it is hovering my mouse on a tab nail or opening the browser or rewinding and those data points are being used to train the algorithm.
Tom Wainwright, Media Editor
The Economist
And so with this data, they know what people like, they therefore know what to commission. The availability of Big Data sets have made it much easier for creators of all kinds to determine what kind of content is going to be popular but AI has really turbocharged that effort.
And that’s what one new company is hoping its use of AI will do to the film industry. Cinelytic is based in LA, the traditional home of film production. But it offers a very modern approach to doing business. Using AI to inform decisions about many aspects of movie making. The ultimate goal is an accurate prediction of how well a film will do, enabling producers and studios to tailor their budgets to expected returns and reduce the risk of losing money.
Tobias Queisser, Co-founder & CEO
Cinelytic
We identify 19 variables that really drive from performance and move the needle. Within 20 seconds the AI is able to provide you with a very detailed risk rated forecast for box office, home video and TV numbers. The beauty of it is you can go back, change any input and within seconds you have a new forecast.
Let’s start with some key decisions, choosing the cast and crew. The AI draws on a database of nearly 600,000 people and offers different predictions depending on who is selected.
Tobias Queisser, Co-founder & CEO
Cinelytic
Let’s swap out Tom Cruise with Matt Damon, also a very strong, well recognised actor that did an action franchise as well that was very popular internationally, we can rerun this forecast, leaving everything else constant. We can now assess every talent out there, actor, writer, director, producer, we created algorithms that figure out the real economic value of this particular talent in relation to the particular content projects that we have in mind.
The AI also takes other factors into account, including a film’s age rating, genre and budget. And thanks to the fast pace of technological change, there’s now another variable in the mix, on that Tobias thinks could be the single most important of all, the script.
Tobias Queisser, Co-founder & CEO
Cinelytic
What that means is, breaking down a script in certain patterns and certain archetypes, characters and really figuring out how these stories could connect to humans.
Cinelytic says its predictions are 85% accurate but stresses there are always outliers.
Tobias Queisser, Co-founder & CEO
Cinelytic
Our tool can really help you get a much better understanding of how your film will perform in the future. Cinelytic is not built to predict outliers.
So it’s time to put its powers to the test. Before the release of Barbie, we asked the multimillion dollar question, how much would it make at the box office?
Tobias Queisser, Co-founder & CEO
Cinelytic
Margot Robbie has been very strong, 2018, 19, 20, then maybe over the last couple of months she really had a bit of a rougher patch and the question now with Barbie, will she bring in the crowds? We’re now talking 700 million for this film in theatres, home video, TV over ten years.
[singing] I got money in my pocket but…
[news report] Barbie has joined the One Billion Club.
Just seventeen days after the film was released.
Two months after the release of Barbie we caught up with Tobias again and asked about Barbie’s phenomenal success and why the AI did not predict the full extent of this.
Tobias Queisser, Co-founder & CEO
Cinelytic
Our tool actually did exactly what it was supposed to do, it helped to show that Barbie will be a successful film would be a positive green light. This is not a crystal ball. We did not foresee that it would become a cultural phenomenon. Barbie is a once in a decade outlier.
It’s a reminder that AI and Big Data do not yet have all the answers.
Tom Wainwright, Media Editor
The Economist
I think there’s been a bit of a turn back towards thinking that actually, human judgement has its place
Competition can be stiff for companies in the entertainment business, especially for streaming services which have proliferated in recent years.
Tom Wainwright, Media Editor
The Economist
Subscriber numbers are plateauing. At the same time, companies are finding that they’re having to keep spending billions upon billions on new content to keep people subscribed.
Enter Stage Left, generative AI. A new plot twist for the industry and a new world of possibilities for creating content.
Tom Wainwright, Media Editor
The Economist
Generative AI allows people to actually create things. They got onto the radar of most ordinary people in about 2022 when apps like ChatGPT and Midjourney came along, allowing ordinary people to register for free and experience some of the seemingly magic powers of generative AI.
Generative AI works by training deep neural networks on large data sets of, for example, images and text. The system learns to identify patterns and when prompted it can generate new content. This new form of AI is enabling some artists to create new forms of art. People like beatboxer, Harry Yeff, also known as REEPS100. Beatboxers use their own voices to make percussive sounds.
Harry Yeff
AI artist and technologist
Through my music and through my art form, I spend a lot of my life trying to sound like a machine.
Using a data set of Harry’s vocalisations, an AI system has generated these percussive noises, a voice that Harry calls his ‘Second Self’.
[music]
Harry Yeff
AI artist and technologist
What these synthetic voices were generated I realised that now a machine was trying to sound like me. The aim was to produce an interaction, almost like a vocal chess match, with some of this synthetic voice. This phrase, where I would normally [beatbox noises] is gone to [noises] and it’s added extra ghost notes or new phrases [music]. My sort of sonic lexicon started to grow. What comes out the other side are streams and streams and streams of this organic Second Self, something that is me but not me. [music] The idea of automating has been a key part in music culture. The newness of it is that identity can be automated.
Harry has also used AI for a ground-breaking collaboration with the likes of ballet in Germany. The dancers’ movements were recorded and this was used to create a data set. AI then suggested new movements which the dancers incorporated into their performance on stage.
Harry Yeff
AI artist and technologist
For the choreography we were able to have video and images, so that becomes the data set, so you’re then able to start producing generative outcomes which are new movements, new emotions.
Harry hopes it’s a sign that AI will open up new opportunities in the Arts to those from less advantaged backgrounds.
Harry Yeff
AI artist and technologist
And the areas that I grew up, where you have no access to resources or anything, you can now with the click of a button have access to intelligence on tap.
Although generative AI promises fresh opportunities and creations, debate will continue about whether it remains more derivative than innovative AND whether it will deliver quantity over quality.
Tom Wainwright, Media Editor
The Economist
I think the real worry about generative AI is that it is essentially parasitic, it sucks up stuff made by humans and remixes it and churns it out again and up to a point that’s great but I think at some point you do need something genuinely new and I think the jury is still out really on whether generative AI is capable of producing something truly original or whether that still needs a human.
Neta Alexander, Assistant Professor of Film and Media Studies
Colgate University
I think we should ask who is going to consume all of these audio segments and videos and I think that what we need to pay attention to is the level of content fatigue on the human side.
Of all the concerns around AI’s impact on the entertainment business perhaps the biggest relates to human jobs. AI is seen as a huge threat by many current workers.
Tom Wainwright, Media Editor
The Economist
With generative AI one can create characters at the click of a button and in a fraction of the time that it would have taken VFX artists to do the same job.
Generative AI is already rewriting the script for how movies are made, from special effects to lip syncing for foreign films and the de-ageing of actors. The disruptive power of AI is one reason why in 2023 writers and actors swapped delivering lines for picket lines.
Newsreader: Hollywood actors are on strike for the first time in 43 years.
Tom Wainwright, Media Editor
The Economist
The strikes that we’ve seen recently in Hollywood by writers and actors are mostly about how they get paid but part of the strikes also is about AI and I think that’s partly what’s helped to capture the public imagination here, the idea of computers and robots coming for the jobs of writers and actors is something that everybody can understand.
Neta Alexander, Assistant Professor of Film and Media Studies
Colgate University
These workers are very canaries in the coal mine, they are telling us what the other industries are going to face. We have the capacity to scan the body, voice and the facial expression of any given actor or actress and then you can use this digital avatar in future projects.
I’m going to show you some magic.
AI driven accounts impersonating the big stars are already popular.
I don’t know why I need to speak Spanish.
But it’s those who are less well-known, both in front of and behind the camera who may have most to fear.
Marcus Hutton
Actor and voiceover artist
Welcome to the last human voice podcast, exploring the moral, legal and ethical issues in artificial intelligence in the creative sectors and in performance…
British actor, Marcus Hutton, believes AI could mean the final act for much of his profession.
Marcus Hutton
Actor and voiceover artist
I’m certainly not anti any AI tools that make workflows simpler or easier. I am uncomfortable about AI systems that are pretty much designed to wipe out a workforce.
One group of workers under threat are voiceover artists. Skills built up over a lifetime can be copied at the click of a button using AI. Marcus fears many artists could be facing their curtain call.
Marcus Hutton
Actor and voiceover artist
There’s a whole ecosystem, a whole set of working practices that could be just chucked in the bin, yeah, it feels quite sad for me. How you look, how you speak, these are things that define you and to have it sort of taken away and synthesised, I don’t think we’re ready as a, as a species for that.
Ready or not, the technology is here and Marcus agreed to have his voice cloned to test out the AI’s effectiveness.
Marcus Hutton
Actor and voiceover artist
So I’m gonna give Speechify a go, they’ve got a free trial of “clone your own voice with one minute of speech in a quiet environment.”
And by the way, as the person narrating this film, I’m as interested as he is to see the results.
Marcus Hutton
Actor and voiceover artist
One minute of material is not a lot of training data but it will be interesting to see what happens. “Speechify voice cloning allows you to create a clone of your voice. In the future, we believe everyone will clone their voice so they never have to re-record a voiceover when a script changes.” Let’s see what it sounds like.
“When anyone else looking to elevate your projects with top-notch voiceovers, I highly recommend giving Speechify a trial.”
It was such a small amount of data to be trained on. I can hear it’s me, that’s what’s kind of weird, it’s me and there’s a bit of nuance and stuff in there.
So perhaps us voiceover artists have been granted a brief stay of execution but AI is capable of cloning voices and the threat to jobs is real.
Marcus Hutton
Actor and voiceover artist
It’s very weird the idea that you might be competing with yourself. AI is definitely taking away jobs rapidly. Station announcements, checkout machines, all that kind of stuff, this is the bread and butter work for a lot of voice artists. I am going to delete the files now because I don’t want a digital version of me to exist.
Tom Wainwright, Media Editor
The Economist
I think some people are probably right to be looking over their shoulder to see if the robots are coming for their jobs. The counter argument to that is that the lower cost of producing this content will mean that there’s more abundant stuff, there’s more things being produced and all of that will ultimately mean more people to work on it.
One estimate suggests that generative AI could add between 2.6 and 4.4 trillion dollars to the global economy annually in the coming decades. Within the media and entertainment industry this technology is expected to increase revenue by 60 to 110 billion dollars a year.
As the technology improves, legal questions over copyright promise to get thornier. Some argue that consent should be obtained each time an artist’s work is used by AI.
Marcus Hutton
Actor and voiceover artist
Once your data, your image, has been captured, your voice has been captured, it can be repurposed multiple times and you have no control over what the output will be. We’re talking about what we call active consent. Your consent has to be gained for each piece of work.
Exactly when and how consent should be obtained. is complex and not easy to decide, and it’s an argument that’s already playing out in high profile lawsuits.
Newsreader: Now some of the biggest and best-selling authors in fiction are suing Open AI.
Tom Wainwright, Media Editor
The Economist
In any dispute like this, the real winners are always the lawyers. But AI companies may well argue that what they do constitutes a fair use of that kind of content, rather in the way that if I read a book and it just informs my thinking in a general way, I don’t necessarily have to pay the author of the book if I write something on a related subject. The way in which AI works is that it sucks up so much information that it’s very hard to pinpoint the effect that any one data point has on any of the output and so it may be that they’re able to argue that they don’t really owe people anything for this. It could turn out though that it’s exactly the other way round and they find that have to foot an enormous bill for every piece of information they have ever used in their training.
Just as it is certain that AI will transform the entertainment industry on many levels, regulation of AI is also inevitable. But so far it is unclear what this will look like or how effective it will be.
Neta Alexander, Assistant Professor of Film and Media Studies
Colgate University
I’m not sure that legislation alone would be enough to stop this tsunami of automation and I think we as consumers actually have more power than we normally realise and we do need to tell companies if you are going to use AI then we are not going to give you our credit cards.
Tom Wainwright, Media Editor
The Economist
There’s no doubt that regulation is coming to AI. Who comes up with those rules and when is one of the big questions of the years ahead.
Hi, I’m Tom Wainwright. I’m the Tech and Media Editor of The Economist. If you’d like to watch more about generative AI and the entertainment business, please click on the link opposite and to watch more of our Now & Next series, click on the other link. Thanks again for watching and please don’t forget to subscribe.