Warner Bailey, founder of the Assistants vs. Agents community, announced on LinkedIn that he had turned down partnerships with major AI platforms that wanted access to his audience. He plans to unveil his own tool in the autumn.
“I’ve turned down partnerships with the major AI platforms who want access to the AvA audience, because I have a deep belief that humans should (and will) continue to drive creativity in Entertainment,” he wrote in the post. He directed interested readers to a registration page for early access.
The community around AvA
Assistants vs. Agents started in 2017 as an internal joke among assistants at the WME talent agency, where Bailey worked at the time. The meme page became a community for young people starting their careers in the industry. According to an interview with The Hollywood Reporter, it now serves around 300,000 people, has 163,000 followers on Instagram and lists around 250 job offers on its job board. The community has 135 university ambassadors in eight countries.
Bailey leads a small team of 12 people aged 21 to 24. The weekly show AvA LIVE launched on 8 April 2026. Bailey himself started as an assistant at WME and was furloughed at Live Nation during the pandemic.
On 10 December 2025, Bailey signed with the agency UTA. According to Deadline, AvA works with more than 30 universities and companies across the industry.
Among its partners are Netflix, Meta, Amazon MGM Studios, Red Bull, Warner Bros. Discovery, Disney, Paramount+ and Apple TV+.
Licensing audiences to AI companies
Bailey also framed his move with a second claim. “Instead of training tech that is designed to replace your jobs, we’ve been quietly building tech to empower the next generation,” he wrote.
Earlier, he told The Hollywood Reporter about the technology itself: “I think generative AI is frightening for a number of reasons and the beat at which it’s developed and the lack of guardrails makes me nervous for creatives in this industry, including myself.”
Closed licensing deals show what AI companies buy from the owners of audiences. In February 2024, Reddit signed a deal with Google for access to user posts to train AI models. According to the Columbia Journalism Review, the deal is worth $60 million a year, and Search Engine Land estimated a similar Reddit partnership with OpenAI at around $70 million a year.
Bailey is swimming against this tide and is not selling access to his audience.
Training data and the European AI Act
The question of whose data trains AI models is exactly what European regulation is targeting. The EU AI Act, in Article 53, requires providers of general-purpose AI models to “draw up and make publicly available a sufficiently detailed summary about the content used for training”. The obligation applies from 2 August 2025, and models placed on the market earlier have until 2 August 2027.
The EU AI Office published the template for this summary on 24 July 2025. For data obtained through web scraping, it requires, among other things, a list of the top 10% of domain names used. For owners of audiences and content, this is a yardstick that shows whose material appeared in the training data.
Bailey did not specify exactly what the planned tool will do. He promised to reveal the details only with the announcement.
Comments