Actors and writers in Hollywood are striking for a number of reasons, including pay and working conditions – but also because of concerns about artificial intelligence (AI).
Writers are worried about control over AI in the screenwriting process, while actors have fears about AI replicas of themselves being used without adequate compensation.
If this all sounds like something out of Charlie Brooker’s dystopian series Black Mirror, that’s because it is – Joan Is Awful, an episode from the latest series released earlier this year, features Hollywood actress Salma Hayek fighting against a streaming service when she realises her AI likeness is being used in any way that it chooses.
The looming cloud of AI hangs over all industries, but are writers and actors right to be worried?
‘Legitimate concern, but a bit overblown’
Speaking to the Sky News Daily podcast, Dr Alex Connock, a senior fellow at the University of Oxford’s Said Business School who has written a textbook on the media and AI, said that while tools such as language model ChatGPT are already being used to produce written content elsewhere, Dr Connock said, there hasn’t yet been any “truly” creative writing produced.
“There’s a really good reason for that, which is that the way it functions, is it’s trained on historic documentation that has been fed in the case of ChatGPT from almost the entire internet,” he said. “So it’s learned its behaviours, if you will, it’s learned its language from what’s already been written. The writers are concerned that something with that facility to do that could then replicate other styles of writing and writing preliminary scripts and so forth, which is a legitimate concern.
“The actors, meanwhile, are concerned that some of the contracts, that particularly extras have been given, have included the rights for their faces to be ingested into AI systems… they’re worried that they will then later be replicated in productions without being paid for that. And that’s probably a legitimate concern.”
However, he said there are nuances to both cases.
“So taking the actors first, it’s quite unlikely that the very legally minded Hollywood companies, say Disney, would take the risk of using an actor’s likeness in a production if they hadn’t explicitly got that actor’s permission to be in that production,” Dr Connock said.
“In any case, if you were making synthetic faces using AI, you probably wouldn’t train the system just on one person’s face, but you would train it on a million or even a billion faces. So the ultimate faces that would be used in those projects would not, in fact, be the likeness of a single actor. So I think the concern may be slightly overblown there, even though I can see the actors’ point that they want to be very careful to protect their likeness.
“What [writers are] concerned about is that the studios might, for example, use AI to write preliminary scripts and then ask the writers to polish them, thereby denying them the right as the originator to make really big money on the project. So I think that’s probably a legitimate concern. But again, a bit overblown.
Please use Chrome browser for a more accessible video player
1:00
“One of the things about generative AI is, of course, it’s not really original at all. And if we think that we love to see original content rather than content that directly represents content of the past, it’s quite unlikely that anyone would be truly entertained by content that was was exceptionally derivative, which is what all the outputs of generative AI systems in fact are. So the writers may be a little bit over-concerned.
“And in fact, there’s a school of thought that instead of synthetic content taking over, actually human content will become ever more valuable. And this is happening because generative AI systems that are trained on synthetic content are subject to something called dogfooding, where the quality of the content goes down and down as each generation of synthetic content is trained on the last. And what those systems really need is original human content.
“So if I was trying to reassure the writers, I would say that actually if they’re good writers and they’re writing real human stories that are based on real experiences, that content could become ever more valuable in the future rather than less valuable.”
‘We want some safeguards in place’
TV and theatre writer Lisa Holdsworth, who is chair of the Writers’ Guild of Great Britain, says AI is a “hot-button issue” in the industry.
“We want some safeguards in place to make sure that we’re not replaced by an algorithm into which our our existing work is funnelled and then, like a sausage factory, scripts come out the other end and our job becomes just giving them a polish,” she told the Daily podcast. “Because that’s not creation – that’s theft, in our opinion.
“So there’s a lot coming over the hill. We want to get in front of it. We want to discuss this with companies, production companies and broadcasters and streamers.”
Currently, “very little” AI is being used for writing, Ms Holdsworth said, and it is still a “fairly blunt instrument” in its current form. This means that at the moment, the strikes are about pre-empting potential future problems.
“But the more and more of our existing work that is fed into those algorithms, into that system, the more of a concern and a threat it becomes,” she said. “It’s being used in graphics, visuals on screen – the recent Secret Invasion, the Marvel series, the opening titles were designed by AI and that made some people feel very uncomfortable.”
Ms Holdsworth says it is “laughable” to think AI can replace human writers and produce the same standard of work.
“So there’s part of me that doesn’t feel that it’s a risk in that way. I think what’ll pop out the other end of any AI system will be incredibly run-of-the-mill, incredibly poor, and will need the human writers – eventually someone will be knocking on the door going, ‘please can you help us with this?’ But if in the interim people lose work, lose status and their position in the industry, that obviously is of great concern.”
While quality would diminish, the power lies with “people who do the sums”, she added.
“The people who make the money, look at the spreadsheets, are not creatives for the most part, and that’s been a good relationship for a long time. Some people make money and work out how to pay for all this, and some of us do the creative work.
“Where the balance seems to have shifted in the last few years is the power lies with the non-creative people within the industry. And if they can find a way to save a penny, not have us annoying creative people around demanding rights and wages and copyright and things like that, then it feels like they will. So actually the threat is not necessarily from the technology, the threat is from people who, with that lack of creativity, don’t realise the Pandora’s box that could be opened if we go down this road.”