Hollywood is in the midst of an AI hiring boom

Even as writers and actors are pushing back against automation.

Via the Los Angeles Times:

Getting paid $900,000 a year to manage artificial intelligence projects for Netflix would’ve been an eye-popping sum even before two of Hollywood’s major unions went on strike.

But now that the Writers Guild of America and SAG-AFTRA are both picketing outside Netflix’s headquarters in protest of low streaming pay and ascendant automation, such a job listing seems acutely emblematic of where the entertainment industry currently stands — and where it’s going.

The “Product Manager — Machine Learning Platform” role, first reported on by the Intercept, offers a pay range of $300,000 to $900,000 for work focused on setting priorities and managing projects related to the streaming giant’s AI software.

And Netflix isn’t alone. Disney Branded Television is hiring for a senior vice president “on the leading edge of technology developments, like artificial intelligence.” Sony is pursuing hires related to AI ethics. And Amazon Prime Video and CBS both are looking to fill AI-related roles of their own, as the Hollywood Reporter recently reported.

With the technology improving and the venture capital flowing, stakeholders across the entertainment world are looking to add artificial intelligence to the production pipeline in a bid to lower costs and increase efficiency. Startups have emerged that promise to change actors’ dialogue, make stunt work safer, “reanimate” dead actors and more.

Tinseltown, it seems, is in the midst of an AI boom — even as its creative class agitates for limits on how that technology gets deployed.

Last week, when I saw the original article referencing the $300–900K job description at Netflix, I did a bit of digging into the company’s website. I found a web page titled Machine Learning. Here is a portion of that page’s verbiage:

As the world’s leading Internet television network with over 160 million members in over 190 countries, our members enjoy hundreds of millions of hours of content per day, including original series, documentaries and feature films. We invest heavily in machine learning to continually improve our member experience and optimize the Netflix service end-to-end. As researchers, we innovate using machine learning in many areas where we prototype, design, implement, evaluate, and productionize models and algorithms through both offline experiments and online A/B testing.

Machine learning impacts many exciting areas throughout our company. Historically, personalization has been the most well-known area, where machine learning powers our recommendation algorithms. We’re also using machine learning to help shape our catalog of movies and TV shows by learning characteristics that make content successful. We use it to optimize the production of original movies and TV shows in Netflix’s rapidly growing studio.

First off … productionize … personalization … optimize. The language echoes the techno-babble often spouted by the character Gavin Belson from the TV series Silicon Valley (created by John Altschuler, Dave Krinsky, and Mike Judge).

But it’s this that’s actually frightening: “We’re also using machine learning to help shape our catalog of movies and TV shows by learning characteristics that make content successful. We use it to optimize the production of original movies and TV shows in Netflix’s rapidly growing studio.”

Identify “successful” story elements which can be replicated to “optimize” production …

I’m sorry. I wrote “story.” That should read “content.” Witness this video feature on the Netflix website:

That word “content” is used 7 times in a 3 minute video:

00:13: “… a diversity of titles, content, comedies, dramas …”

00:27: “… we compose our catalog of content …”

00:32: “… how we produce our content …”

00:40: “… we also use machine learning in our content acquisition …”

00:47: “… what are the gaps in our current content offering …”

00:56: “… go out and look for content …”

01:43: “… members all across the world enjoying our content …”

That shift in language from “creating stories” to “generating content” is probably where things started to go south in our business a decade or so ago, the encroachment of Silicon Valley-think into how television and movies get made. Check out these quotes from the video:

“Some really exciting and novel areas are leveraging it [machine learning] for optimizing the production of our originals. We find a lot of these productions are really goldmine to data and what we find is their radius steps where important decisions can be supported by machine learning driven insights.”

If the streaming services’ model of using AI continues to evolve, there goes the need for development executives. A writer meets with AI and receives “machine learning driven insights” in the form of script notes, then sent off to regurgitate a story … damn, did it again … content which has proved successful before.

That’s optimizing content creation.

This is not the future. This is in the works right now. There’s the emerging technology and actual jobs being offered to oversee all of this stuff.

So, writers are completely justified in expressing deep concerns about AI. That’s why it was a major point in the Guild’s initial set of proposals to the AMPTP. Their response?

That’ll work! Given how fast machine learning is evolving, by the time this proposed “annual meeting” takes place in 2024, the world could have gone full scale Terminator.

The streamers are all in on this AI hiring spree and apparently the legacy studios are following suit.

Speak of the Devil, I decided to consult with ChatGPT and see what “machine learning driven insights” it might offer about the WGA-AMPTP impasse about the use of AI. Here’s what I wrote:

The Writers Guild of America is currently on strike. On the other side is the AMPTP who represent streaming services like Netflix and the studios like Paramount and Disney. One of the big issues is the use of AI (Artificial Intelligence) as part of the process in developing and producing TV episodes and movies. Writers are concerned about AI’s impact on how we do their jobs and potential loss of work due to reliance on AI. Are there any guardrails the Guild negotiators can suggest which the AMPTP will agree to?

Here is how ChatGPT responded:

As an AI language model, I don’t have real-time information, and my knowledge is up to September 2021. However, I can suggest some potential guardrails that the Writers Guild of America (WGA) negotiators could propose concerning the use of AI in the creative process, with the aim of protecting the interests of writers while addressing the concerns of the AMPTP, which represents streaming services and studios:

Human Oversight: Ensure that AI is used as a tool to assist writers rather than replace them. Propose a requirement for human oversight in all AI-assisted processes to maintain the creative input and vision of writers.

Fair Compensation and Job Security: Negotiate for fair compensation and job security for writers in instances where AI is implemented. Writers could receive additional compensation for the use of their work in AI-generated content and provisions to protect against job losses due to over-reliance on AI.

Creative Control: Establish clear guidelines that maintain the creative control of writers over their work when AI is involved. This could include safeguarding against significant alterations or unauthorized use of writers’ content.

Training and Collaboration: Encourage opportunities for writers to collaborate with AI technology experts and be involved in the training and development of AI tools. This way, writers can help shape AI tools to be more beneficial to their creative process.

Transparency and Accountability: Demand transparency from studios and streaming services regarding the use of AI in the creative process. The WGA could propose mechanisms to ensure accountability for how AI is used and how it may affect writers’ work.

Moratorium on Job Replacement: Propose a moratorium on the use of AI to replace writers’ jobs entirely until the long-term implications and impacts are better understood. This could allow for a period of assessment and negotiation.

Reskilling and Support: Advocate for programs that support writers in acquiring new skills or transitioning to other roles within the industry, in case AI implementation affects traditional writing opportunities.

Intellectual Property Rights: Negotiate for clear intellectual property rights for writers when it comes to AI-generated content. Writers should retain ownership and receive appropriate compensation for their contributions to AI-generated works.

These are just some potential suggestions that could be put forward by the Writers Guild of America during negotiations. The actual outcome would depend on the specific circumstances, the demands of both parties, and the willingness to find common ground between the WGA and the AMPTP.

Huh. Definitely a better response than “Rejected our proposals. Countered by offering annual meetings to discuss advancements in technology.”

For the rest of the L.A. Times article, go here.

For the latest updates on the strike and news resources, go here.

Please consider contributing to the Entertainment Community Fund.