Brainstorm AI London 2024: Responding To AI’s Disruption Of Entertainment And The Arts
Josh Berger CBE, Founder and Chairman, Battersea Entertainment Lynda Rooke, Actor; President, Equity Alexandra Shannon, Head of Strategic Development, Creative Artists Agency Moderator: Ellie Austin, Deputy Editorial Director, Fortune Live Media, FORTUNE; Co-chair, Fortune Brainstorm AI London
Category
🤖
TechTranscript
00:00Hello again everyone, and hello Josh, Linda and Alex.
00:04Thank you so much for being here.
00:05Alex actually has just flown in this morning from New York,
00:07so top marks to her for being here.
00:11Linda, I want to start with you now.
00:13So, equity represents a whole range of artists.
00:15We were talking backstage, you said,
00:17in addition to actors, musicians, wrestlers,
00:20models, radio stars, which AI use case
00:25do you think will be most disruptive
00:27to the members of equity?
00:28Probably screen, and to some extent life as well,
00:32live theatre, because of streaming of shows, et cetera,
00:35that suddenly come in since the pandemic
00:37or during the pandemic.
00:38But particularly film and television,
00:42as well as voice as well, yeah.
00:43Voice as well, the replication of voice.
00:46And I know that you're campaigning quite hard
00:48at the moment to raise aware of these issues,
00:49because there are no formal regulations in place.
00:53What would you, what guardrails would you like to see
00:56around AI in the arts industry?
00:59Well, there's a number of issues.
01:00It's to do with transparency for a start-off,
01:03knowing where your image is, what the rights are,
01:07where it's being used, et cetera.
01:09Consent as well, and control over that image.
01:12You know, that you have the right
01:13to withdraw that consent, et cetera.
01:15And the biggie, remuneration.
01:18Basically, if you're going to use our image
01:21and our voice, et cetera, we want to be paid
01:23and we want to know about it,
01:24and we want to be paid properly.
01:26We're going to come back onto
01:27what properly means in a second.
01:29Josh, now you've worked in the media, TV, film,
01:33theater industry for decades,
01:34and you were at Warner Brothers,
01:36where one of the things you did
01:37was you ran the global Harry Potter franchise,
01:39which is something very close to my heart,
01:41and I'm sure many millennials in the audience as well.
01:45I want to focus on the films for a second.
01:46If that first Harry Potter film
01:48was coming out in the AI age,
01:50how would its production be different,
01:52and how would the end product look different?
01:56That's a good question.
01:56I mean, the first film,
01:59even as compared to the last films,
02:02there was a kind of revolution
02:04in the visual effects industry in this country as well,
02:08and Potter kind of brought on what is now
02:12the world-leading VFX industry based here.
02:17And so, in some ways, it was a step in that direction
02:22because they were computer-generated images
02:25in the latter films,
02:27and in the earlier films, not nearly as much.
02:31The castle was, for example,
02:34it was of an annex, for example.
02:36And then you'll see that there was much more digital imagery
02:40in the later films as technology advanced.
02:43Well, in the AI era,
02:46it would be possible then to do more of that,
02:49and it might look even more real,
02:53and it would be created with a verbal prompt in 60 seconds
03:01instead of many, many people working to create it.
03:05Now, again, that presupposes
03:08that we've worked out a lot of these issues
03:11about remuneration for people
03:12who would normally be doing that work,
03:15humans doing that work as opposed to a machine.
03:17And then, of course, the machine
03:19that would create the visuals
03:21would be building them based on scraping
03:24all of the visuals that have already been created
03:27by somebody.
03:28So, again, the remuneration for the work
03:32and the attribution of that work
03:33because the transparency is really important.
03:36And then, again, I don't suspect anybody knows
03:40where every single image and every piece of text
03:43was ever created because that's what's out there,
03:45and that's what's being scraped,
03:46and that's what's creating the foundation of the AI,
03:50which then creates visuals and text and scripts
03:53and everything else.
03:55Now, Alex, at CAA, you've come up
03:57with a pretty interesting and innovative way
04:00to manage the thorny issue of digital doubles,
04:03and you've created something called the CAA Vault,
04:05which allows your clients, your actors,
04:07to create their own digital doubles
04:10and then have control over how and when they're used
04:12and be paid for them.
04:13Can you talk us a bit through how that came about?
04:16Sure.
04:17So, it's no surprise to anyone in this room
04:19that a few years ago, we noticed that on the one hand,
04:22there was just rampant misuse of our client's name,
04:25image, likeness, voice, body of work without their consent.
04:28And on the other hand,
04:29we were noticing interesting opportunities
04:32in companies who wanted to work with our clients
04:34and their digital rights,
04:35but didn't have a way to do that
04:38in an ethical, talent-based mechanism.
04:41And so we decided to launch what we call the CAA Vault.
04:44It is, on the one hand,
04:45we are proactively creating our talent's digital likeness.
04:49We are scanning their image, we're scanning their voice,
04:53we're scanning their likeness,
04:55and we are then storing that on their behalf.
04:58The purpose of this is, on the one hand,
05:01we know that the law is going to take time to catch up.
05:04And so this is a mechanism for our clients
05:07to actually create and own and have permissions
05:10around their digital identities.
05:12And we think, over time,
05:13that will actually help the law catch up faster,
05:16because now there is a legitimate way
05:18to work with one of our clients.
05:19There is a legitimate way to compensate them
05:21and to give them credit for their work.
05:24And on the opportunity side,
05:26for those who are interested, and not all are,
05:28but for those who are interested,
05:30this provides a way for us to help set a precedent
05:33for anyone who wants to work with one of our clients
05:36and their digital identity.
05:38There is a mechanism to have them be compensated.
05:40There is a mechanism to work with them
05:43in the right way and unlock opportunities.
05:44Have you had to convince your clients
05:46to take advantage of this?
05:47Are some people skeptical?
05:48What's been the take up?
05:49Absolutely, it runs the gamut.
05:52On the one hand, you have clients
05:53who are very skeptical of this space,
05:56and on the other side, you have those
05:57who are quite interested in thinking about
05:59how to take advantage of this space.
06:01And the clients that we're scanning
06:02are actors and actresses, but it is also musicians
06:06and athletes and digital influencers and politicians.
06:09It runs the gamut,
06:11because we think that by starting
06:13and starting to set the precedent,
06:15that we think this will ultimately benefit everybody.
06:18We all know sort of the conversations
06:20we were talking as we were walking in
06:21about some of the conversations
06:23over the last year or so.
06:25We can't just stop at the top in any field.
06:28It has to be a place where we can set the precedent
06:32that everyone benefits from.
06:33One more question to you, and then Linda and Josh,
06:35I'm gonna bring you in on this subject.
06:37You talked about remuneration.
06:38How is it calculated if one of these digital doubles
06:41is used by a studio?
06:42What are the metrics for deciding
06:44what the client should be paid for that use?
06:46It's a great question, and the market is forming,
06:49to be honest.
06:50It is happening in real time.
06:52But the idea is that if you're to work
06:54with somebody's digital double,
06:56you aren't working with that because you think
06:58you can work for that person in a more cheap way,
07:00that it's creating some big cost efficiency for you.
07:03At the end of the day, if you are working with somebody,
07:06the value of that person to a brand, for example,
07:10if it's a brand endorsement deal,
07:11the value is still in that person representing your brand,
07:14whether it is them fully in person and performing,
07:18whether it is using some of these AI tools
07:20powered by a digital double to do post-production fixes
07:24or to help enhance the value,
07:26to think about localization of content, for example.
07:29Okay, so it's more a scheduling or convenience tool
07:34rather than a cost-cutting measure.
07:36Exactly.
07:36Linda, Josh, what do you make of this?
07:38Linda, would you be encouraging your members
07:40to get involved?
07:41I think if you have status or profile in the industry,
07:45there's probably a very good reason
07:46to sort all that side of things out,
07:48and I'm sure it's going on.
07:50But I represent 50,000 members,
07:53not of all whom you will know who they are, et cetera,
07:56some just starting out in the industry.
07:59And they won't have the power, et cetera,
08:01to be able to work through those problems as well,
08:06or the financial wherewithal to do that.
08:10So those are the issues around making sure
08:12that everybody's protected,
08:13added to which, dare I say,
08:15the other people who are not in the room
08:17to be able to defend themselves
08:18are those who are no longer with us,
08:20where somebody's image and voice, et cetera,
08:23who is dead, who is then being brought back to life
08:27without the consent, without their consent.
08:30So it's our estates trying to manage that.
08:33Well, yes, but there's also the other thing, of course,
08:35when a lot of programs were made in the 70s, 80s, 90s,
08:40what was going on, there was no AI.
08:42So within what we have is called collective agreements,
08:45there was nothing to control
08:46the use of those images and rights.
08:49Because back then, as you know,
08:50we were dealing with CDs and cassettes,
08:52and everybody's cell phone or mobile phone
08:54was the size of a brick.
08:56So things have changed now.
08:58And unfortunately, relying on the agreements
09:01that were there in place and the contracts that were signed,
09:04something has to be updated.
09:05So are you speaking to the government about this?
09:07Yes, yes, there's a number of fronts on this one.
09:10There's the intellectual property side of things,
09:13and there's also GDPR as well.
09:16There's a number of routes.
09:17Plus, of course, if they could adopt 1B,
09:19Option 1B of the Beijing Treaty,
09:22that would be useful, because so much of our work
09:25in English-speaking countries, we work a lot with the USA.
09:28What would that entail?
09:29I'm ignorant to this.
09:30Well, basically, we work with our collective agreements,
09:34because we obviously, all the time,
09:36the USA and Canada, film over here, et cetera,
09:40our actors work over there,
09:41and there's a kind of multiplicity around that.
09:43And it would just ease the problems
09:46around property rights, et cetera.
09:48Okay.
09:50Josh, you were saying backstage that you were in LA
09:52around the time that there were
09:53the SAG-AFRA strikes last year,
09:56which those strikes were about the streaming economy.
09:58They were also about a then lack of regulation around AI.
10:02And on the whole, it seems that humans won, as it were.
10:04There's now guardrails around what can happen with AI
10:07in a writer's room, and also,
10:08when it comes to talents doubles.
10:11How do you feel about the outcome of the strikes?
10:13And maybe, what could we learn from them here in the UK?
10:17First of all, I think it's too early to know
10:19what the outcome is.
10:20I think the headline sounded good
10:23for the actors and for the writers.
10:25And it was good that the studios finally capitulated
10:29and did something, because for too long,
10:34they were immovable on the subject.
10:36And it cost a lot of people a lot of work
10:40and a lot of money that they desperately needed,
10:42a lot more than the studios who were putting them
10:44in that position.
10:45And honestly, they could have gotten to that deal
10:49months faster, in my view.
10:51But aside from that, I still think it's too early,
10:55because if you calculate also the cost of the strike
10:59in lost business for everybody.
11:02Warner Brothers Discovery restated their numbers
11:05based on the strike impact, and the numbers were horrendous.
11:10And again, if they had worked that into
11:12their negotiating position, again,
11:15David Zaslav would have pushed a lot faster, I think.
11:17To resolve.
11:18I would have thought so.
11:23So I think it's too early to do the cost-benefit.
11:25And it's sort of a step in the direction
11:29of not resolution, but just some modus vivendi
11:32while we see what happens with AI
11:35and how AI is going to be used.
11:38So more positive than negative, because there was a deal.
11:43I mean, the real negative case was just that kind of
11:45standoff for all that time.
11:47And there was an acknowledgement among the studios
11:49that, for example, if you write something with AI,
11:53you can't, you know, AI can't be the credited writer
11:55of anything, it's gotta be a human being.
11:58There's just interesting data, most recently,
12:00just with CEOs, I mean, half of CEOs now are using AI
12:05and passing off that work as their own,
12:08which you probably saw just in the last couple of days.
12:11So, I mean, the fact is this tool is going to be used
12:14by many, many, many people, including in our industry.
12:18And we need to figure out how to, again,
12:20how to pay people properly.
12:22There's numerous court cases at the moment
12:24between big tech companies and the New York Times
12:26or the Authors Guild over the fact that
12:29original source material that's used to train the AI
12:32is never credited.
12:32So I guess it's also an IP law question as well.
12:36I'm gonna open it up to the audience now
12:39to see if anyone has a question for our panel.
12:42If you do, please raise your hand and we will come to you.
12:45If not, I will continue with my questions for now.
12:49Okay.
12:51Oh, no, we do have one question.
12:52And please state your name and the company that you're from.
12:55Thank you.
12:56Hello, thank you.
12:56My name's Jack Aldane.
12:58I'm a podcast producer for Global Government Forum.
13:01I just wanna sort of pitch three scenarios
13:05and ask the panel how plausible
13:07they think they will be looking forward into the future.
13:10With the ability to take the likenesses of celebrities
13:14deceased, though it comes with ethical problems,
13:17or still alive, and create kind of 3D models of them,
13:21how plausible is it that we'll see, say,
13:23the Beatles headlining Glastonbury in 2035?
13:27Second scenario, could actresses and actresses,
13:31based on the sort of work that you're doing with them
13:33at the moment, potentially work forever?
13:37And if that's the case, well, I guess like children
13:40really would never have to work again.
13:41But the point being, could that actually be,
13:44is that plausible if the value that they put on their work
13:46isn't that they do it in person,
13:48but that they simply bring their likeness
13:50and their data to it?
13:52And I guess the third would be,
13:53do you think it's plausible that public figures
13:55may want to take advantage of that, too?
13:57Might we live one day in a world
13:59of immortal presidents and prime ministers?
14:02Okay, why don't we?
14:04It's a scary thought.
14:06God help us. Sounds horrific.
14:08Let's have each of you take one of them.
14:11So Alex, maybe if you start with the Beatles
14:13headlining Glastonbury question.
14:15Yeah, so they're fascinating questions,
14:18and I think these technologies
14:20are only going to continue leading
14:22to this explosion of creativity.
14:24I think we're only scratching the surface
14:26for what's possible in the long run.
14:29Certainly thinking through how the technology is applied
14:32and thinking through the human-centric element is paramount.
14:36But I think that world where you think about
14:40the legacy of somebody, you think about their ability
14:42to continue making impact years and years and years to come,
14:46I do think is an interesting one.
14:48I equate it very much to when digital cameras
14:51were coming out.
14:52It's my favorite analogy with this space
14:53because that didn't get rid of the traditional art form
14:57of painting and the traditional in-person artwork.
15:01It just created this whole new incremental set of work,
15:05and it sort of expanded the pie in many ways
15:07and grew the pie.
15:08And so I very much think that,
15:10is there a scenario such as the Beatles example?
15:13I'm not sure.
15:14We're seeing versions of that with even here in the UK
15:16with ABBA Voyage, which I think is a fascinating one
15:19where if you think about the dynamics of a group of people
15:23that otherwise may not be comfortable traveling,
15:26may not all be able to be in the same place,
15:28they're still able to reach fans and engage with fans
15:31in the right way, but they were in control of that.
15:34And it's a model that puts them front and center.
15:37I think those sorts of experiences
15:39and ways to continue reaching fans for generations to come,
15:43I do think is a powerful opportunity.
15:46And then Linda, briefly on the question
15:47of actors and actresses working forever.
15:50Yeah, I mean, I understand where you're coming from
15:52with that kind of thing, but the truth is,
15:55this is an industry that demands change
15:58and innovation on a constant basis.
16:01There's always a fresh new face, whatever,
16:03being looked out for.
16:05So whilst there is a potential, I guess,
16:07for somebody to be playing the Harry Potter kid
16:10forever and a day, the point being is that in the end,
16:13that you, the audiences, will demand different faces,
16:18changes to those things.
16:19We will move on, you know.
16:21Laurence Olivier is no longer with us,
16:24we've enjoyed looking at his films, et cetera,
16:26but we want fresh faces, we want different,
16:28we want other younger people coming through.
16:31There could be the chance of, you know, you never age.
16:34You know, the Peter Pan effect,
16:35that'd be an interesting one.
16:36But overall, I think it's in our collective agreements,
16:40write it in there, pay as well,
16:42and send it on to the families, et cetera.
16:44But in the end, you, the audience,
16:46will tire of that repetition.
16:48Trust me.
16:50I want to end on a more upbeat note.
16:53Has anyone here got an example in mind
16:55of a piece of art that's been created using AI
16:57that they think is spectacular
16:59and wouldn't otherwise have come to be?
17:03Yes.
17:03Josh, tell us.
17:05I thought you were gonna ask me that question
17:06about the politicians.
17:07I mean, can I just answer it very quickly?
17:09Yeah, please do.
17:11If the questioner asked that question of Donald Trump,
17:13he'd say, yes, I want to live and be the president forever.
17:16That was the first thought,
17:17and that's like the biggest nightmare in history.
17:20So, said as an American.
17:25What was the other question?
17:26The way AI has been used for good in art.
17:31Damien Hirst did an NFT of something called The Currency,
17:36which was really cool, and his machine created,
17:39but he did another series.
17:41I forget what it's called.
17:42The Spun Paintings, or the Spun Colors.
17:45I mean, they're really, really beautiful, and they're AI.
17:47And you can kind of build your own, actually,
17:50and sort of tell the machine what you want,
17:53and you can see, and he puts his name on it,
17:55as he does, and he's good at that.
17:57But they really are stunning, actually,
18:00and that's AI art, and there's a marketplace now
18:03for that AI art in the NFT space.
18:07Josh, Alex, Linda, that's all we've got time for.
18:09It's been a joy speaking to you.
18:11Thank you so much for joining us.
18:13Thank you.
18:14Thank you.
18:14Thank you.
18:15If you go over here, I've got that.
18:16Enjoy the next one.