Calls to innovate in AI could potentially endanger the creative industry, since there may be too few regulatory guardrails on companies, says the lead creative industry negotiator from Hollywood
ADVERTISEMENTEU politicians should remain on guard for the impact of the EU AI Act since its effects remain to be seen, the key negotiator behind a 148-day strike by Hollywood writers whose effects rippled beyond California warned in Brussels yesterday (3 October).
Ellen Stutzman, executive director of the Writers Guild of America West (WGA), led negotiations on behalf of 11,500 screenwriters against producers of series and films united in the Alliance of Motion Picture Producers (AMPTP) in the industrial action, which fought to protect creative work from the use of large language models like ChatGPT.
She spoke to Euronews on the fringes of an event organised by the Friedrich-Ebert-Stiftung and trade union UNI Europa – led the WGA during the labour dispute.
“I think it’s a good development and a starting point, to push for better transparency, to get consent and how to deal with remuneration,” Stutzman told Euronews of the EU AI Act, but she said that «the knock-on effect of AI rules in Europe and elsewhere remains to be seen».
Stutzman said that in the US and EU there’s often the argument that AI should be deployed without the risk of stifling innovation.
«I think it [this argument] can lead regulators and politicians to think we shouldn’t do anything,» she said, adding: «But this would be incorrect. No one asked companies to take over everyone’s work and create these models.»
After the Covid pandemic, production companies applied budget cuts and writers were being asked to do more work with a smaller staff and less pay.
Fearing redundancy, the writers went on strike, demanding that AI tools be used only to assist, rather than replace, research or script ideas.
The strike had knock-on effects on European productions and collaborations involving US studios and performers, causing delays and cancellations.
They finally concluded a deal which recognised that AI-generated material should not be treated on a par with human-written text and that writers could not be compelled to use it.
Traditional media companies like Disney, Sony and Netflix are not the ones that have created large language models, according to Stutzman, «so, we’re able to negotiate with them about the terms and conditions of employment for our members», she said.
«But we don’t get to speak to OpenAI; there remains an ongoing need to deal with how these AI systems have used copyrighted works,” she added.
Workplace and the AI Act
The EU’s AI Act – which regulates AI models according to the risk they pose for society – entered into force in August after long negotiations between the European Parliament, the 27 member states and EU Commission. The first provisions will only take effect as of six months’ time, but the Parliament is not sitting still.
Lawmaker Brando Benifei (Italy/S&D), who was the Parliament’s co-rapporteur on the Act, has repeatedly called for more specific rules on AI and the workplace.
Benifei, who spoke at the same event in Brussels, said that the AI Act already foresees protecting original content through an obligation to watermark AI-generated content.
ADVERTISEMENTThe Parliament will set up an AI Act monitoring group to keep track of the developments and the actions in the member states and by the Commission. It is expected to be finalised mid-October.
The chair of the Parliament’s Employment and Social Affairs Committee (EMPL), Li Andersson, said in August that “looking at AI and working life is very important” due to the rapid changes. “We need to make sure that both EU and national policymakers keep up with the pace at which things are changing,” she said.
Earlier this year the Commission asked an external agency to examine the current situation on use of AI, which could possibly prepare the ground for a fresh policy proposal in the new mandate.
