Cooler Heads to Prevail Over Generative AI

Robots On Laptops

The creative apocalypse is unlikely to unfold as the calculating minds of media executives on corporate boards think through the risks of Generative AI.

Generative AI is the trending term for those artificial intelligence systems that are fast becoming the number one topic of conversation in board rooms across the Media & Entertainment industry.  This is driven in part by the algorithms of Google and Bing, which are currently directing billions of internet users to search results that promise fast money for adopting Generative AI solutions offered for everything from knitting patterns to script writing. 

If you want your laptop, tablet or phone to generate that novel or movie script you always wanted to write, Generative AI is there to do it for you.  ChatGPT, from Open.ai, will even do it for you gratis.  That is, free.  In other words, for no payment.

It seems that Generative AI has captured the imaginations, aspirations and ambitions of nearly everyone paying attention to what the internet is telling us.  Google itself lists the search terms, “AI” and “ChatGPT” as among the top trending search terms in the United States since January, when ChatGPT erupted into the popular imagination. 

See, the Google Trends comparison

In some respects, Generative AI represents a tool that humans can use to craft artistic expression faster, if not better, than they could without it.  Some may argue that Generative AI is analogous to the steam engine, the loom or even the printing press; each of which ignited a new era for human-kind.  Objectively this may be true.  But nothing one reads or listens to about Generative AI seems to be objective.  Two camps are emerging, both being kindled and fed by promoters of all kinds with as much interest in the argument itself as the outcome. 


You may also be interested in …

The Media Big Top

Entertainment Grows as Private Equity Pours Money into Disruption

PE Investors are fast becoming the number one source of capital for a US$2.4 trillion industry struggling to cope with unstoppable growth.


Proponents of Generative AI see the utility of it, and embrace it’s use as another instrument in the creative process that drives modern life and, particularly, commerce.  Opposition is focused on the limitations of AI itself, and the human capability to disregard those limitations.  The outcome being rapid adoption of the technology to the detriment of both the quality and the quantity of jobs that require some level of literacy and the ability to string words together or hum a tune.

The Media & Entertainment industry relies almost exclusively on creative workers to work at creating what Media & Entertainment companies sell to audiences.  According to PWC’s most recent industry analysis, selling entertainment content to audiences is likely to generate US$2.4 trillion in revenues this year alone.

The anxiety that arises from job insecurity is potent, particularly when considered on a macro-economic level.  The creative, but often lower to middle-income employees who feed their families through the written word, the image or the musical note have only to look at layoffs from the booming tech industry to imagine what next year might bring if Generative AI takes over. 

Most will choose to save their money, rather than spend on a car, a new house or all of the other things a consumer economy needs people to buy.

Corporate Care

Creatives that generate the entertainment product sold by the largest Media companies may be lucky enough to dodge this AI generated bullet.  This will not be due to concern for worker welfare, but out of concern for the legal basis upon which the Media & Entertainment industry generates money.

This is being demonstrated in board rooms across the industry, as discussion of Generative AI begins to coalesce around the mundane attributes of intellectual property rights and contracts.  Particularly, licenses from copyright owners. 

That would be the creatives.


“What the artist creates is more than simply generated content; it is intellectual property. Specifically, the artist creates a copyright.


The Wall Street Journey recently reported on such a discussion at Paramount Global.

Phil Wise, Paramount Global’s technology chief, recently rang the alarm bells during a presentation with company executives, including Chief Executive Bob Bakish.

According to the WSJ, Wise used the Open.ai’s other free to use AI for images, DALL-E, to generate several of Paramount’s most iconic entertainment characters, including SpongeBob SquarePants and Optimus Prime.

“This woke everyone up.” Wise told the Wall Street Journal. 

But not for the reasons many might think.

Credit Where Credit’s Due

Revenues generated by the Media & Entertainment industry are not paid directly to the artists who create the entertainment content audiences spend their money and leisure time on.  Rather, the rights to generate money from entertainment content is sold by the artist to distributors, who then license exhibitors, television channels and streamers to charge money to show it. 

For the most part, those distributors, streamers and television channels are all companies within the same media conglomerate, such as Paramount Global. 

If they can, someone in one of those companies will charge advertisers to pitch their goods and services to whoever has been gathered to watch, listen to or experience the entertainment that the artists created. 

What the artist creates is more than simply generated content; it is intellectual property.  Specifically, the artist creates a copyright.  Every novel, every script, every piece of music and every computer programme created by a human is intellectual property originally owned by the creator as copyright.

Every penny that is generated from every movie, every television show, novel, song or video game has been carefully protected by legal agreements transferring the right to do so from the artist to the distributor and then to the exhibitor. 

If the creator of the art is not human, then there is no copyright, and thus no right to earn money from that art.

The Penny Drops

To make things even less simple, Generative AI systems tend to be built on foundational models of artificial intelligence.  These models are “trained” on the expressions and combinations of words, images and sounds found primarily in data available publicly over the internet.  The “GPT” in ChatGPT is an acronym for “Generative Pre-trained Transformer”.  It “transforms” a query or prompt entered in the format of a chat box into a response generated from topical data “trained” into the programme from the world wide web.

So, effectively, everything that comes out of ChatGPT is a paraphrase of what someone else published on the internet.  Sometimes, the response is not a paraphrase, but directly plagiarised material that may, or may not be used by ChatGPT within its intended context.

What ChatGPT, and other Generative AI systems generate in response to a human user is not technically copyright property in the traditional sense.  At best, the material generated can be attributed to the human who “prompted” the material from the AI. 

But this is not settled law anywhere.

At worse, the Generative AI has violated the copyright protections of the human creatives whose works, available on the internet, were used to “train” the AI.  Somewhere in the middle is the legal position that no one owns the copyright to what Generative AI has created.

This is a good reason for Open.ai not to charge any fees for the material that its Generative AI produces.

It is also reason for the Media Majors to be very, very worried.

“One of the biggest risks here is that these engines can generate our intellectual property in new ways, and that is out in the hands of the public.” Wiser told the WSJ.

What the WSJ didn’t report is why this is such a big risk. 

On February 3, 2023, Getty Images, one of the world’s foremost providers of stock images on the internet, sued Stability AI, one of the fastest growing Generative AI startup companies.  Getty Images accuses the startup of using over 12 million of its photos to “train” its popular image generator, Stable Diffusion, without license from Getty Images, or payment of any fees.

In the formal Complaint filed with US District Court in Delaware, Getty Images stated that, “Rather than negotiate a license with Getty Images for the use of its content [ … ] Stability AI has copied at least 12 million copyrighted images from Getty Image’s websites, along with associated text and metadata, in order to train its Stable Diffusion model”. 

Getty Images has asked for a jury trial and is asking the court to award damages, including a potential award of US$150,000 for each infringed work.  That’s potentially sixteen figures.  If that’s not a message to Boards of Directors everywhere, nothing will be.

Imagine if an artist sold licenses to Paramount for a new animated series based on characters generated by Stability AI’s Stable Diffusion, or Open.ai’s DALL-E.

Read the full Complaint from Getty Images here.

Time and Place

Still, there may well be a time and a place for Generative AI.  Just as there was a time and place for the cotton gin or even the wheel. 

AI is essential in the production of high-resolution graphic images used increasingly in real-time digital effects on film and television projects. Using immersive, virtual stage “volumes” software engineers and film directors use AI to integrate these into in-camera VFX as a scene is being shot, reducing time in post-production.  These advances save time and money in the production process just as demand for content is increasing.


You may be interested in …

Digital Previs

Digital Pre-Vis: Ask Damocles

Dear Damocles. Is digital pre-vis important in film-making today or are story boards still good enough?


In the recent Oscar-winning film, “Everything Everywhere All At Once” the producers used a Generative AI tool to delete a wire system with noticeable pullies that manipulated two characters in one of the film’s climactic scenes.  Evan Halleck, the film’s VFX specialist, told the WSJ that this saved a couple of days of tedious work.

Other media executives are more pragmatic when in comes to any choice between human and AI creation.

“AI uses what already exists out there and that doesn’t create innovation.” Jason Blum, founder and CEO of the horror-focused production company Blumhouse, told the WSJ.

“If you look back at the most celebrated movies and shows over time, like ‘Citizen Kane,’ I find it hard to believe that any of those would exist in an AI-driven content world.”  He added.

Perhaps the only creatives who rely on the written word who should be worried are the lawyers.  Surely, ChatGPT can write the documentation required for the next major lawsuit.  There’s one available right there on the internet after all.

Read the Wall Street Journal Article here.

1 Comment

  1. the copyright issue is perhaps even more crucial in the tech world, where patent ownership represents the underlying value of most software. This will probably be fought about in various courts for years. There’s also an interesting defamation case being brought by an Australian mayor who has been falsely accused of fraud by Chat GBT, which – as anyone who has tried it will know – is very often wrong about “facts”.

Leave a Reply

Previous Story

Soft Money, Tax Credits and Production Rebates

Next Story

Hollywood’s Sound Stage Supremacy Under Pressure