Lights, camera, action!
Exactly a year ago, more than a terabyte (that's a lot of data) of chat logs and internal conversations from Disney was leaked.
Among the content that was published online were business sensitive information about future collaborations with gaming companies, new projects, software development, and salary conditions.
The hacker group NullBulge claimed responsibility. But the weird thing was the group hadn't done this for their own financial gain. NullBulge are "hacktivists" who carried out a deliberate and targeted attack against Disney. In security terminology, this is called Spearphishing, a form of attack where individuals, organizations, or businesses are the target.
So why was Disney targeted?
NullBulge stated that the goal was "to protect artists' rights to fair terms and pay." The attack was part of a protest against AI use in art and the film industry. The group took matters into their own hands because of how Disney structures the contracts they enter into with artists, performers, and others. In addition, the hacktivists are concerned about how the Mickey Mouse company uses AI and that Disney, according to NullBulge, "obviously doesn't care about the consumer." Disney animators were worried that their jobs would be replaced —and their work destroyed— because of artificial intelligence. In November 2023, 93 percent of animators voted to become part of The Animation Guild union.
A long strike
Around the same time, screenwriters in Hollywood had enough. A 146-day strike organized by The Writers Guild of America, which became the second-longest strike in Hollywood history, was largely about artificial intelligence. The writers were concerned about AI use and that it endangered their jobs. And by now, you might notice a common thread: fear that AI will replace our jobs.
Though I must add that another technology also played an important role in the writers' strike: Streaming platforms. Most series made today contain somewhere between eight and twelve episodes. This is a significant decrease from the approximately 20 episodes per season that were common when we had "traditional television." Fewer episodes mean lower pay, and so writers wanted to secure themselves through strengthened rights and compensation for reruns.
Trying to find the solution
Star directors Joe and Anthony Russo are behind mega-blockbusters like The Avengers: Endgame. The brothers own the production company Agbo, and now they're working to prevent artificial intelligence from becoming a threat to filmmakers. Their goal is for AI to become a tool the industry can benefit from in creative work, rather than a technology that will replace humans altogether.
We don't know exactly what the directors are working on, but they have previously expressed enthusiasm for transmedia solutions:
Films, games, and series struggle with profitability, and one way to save costs is to use 3D models of characters, props, and sets. These can then be reused in games, films, as part of an amusement park, or on other platforms. This usage is what's called transmedia.
Transmedia might become the dominant form of media going forward, - Anthony Russo
Video generation
Hollywood is struggling. Yet, if Los Angeles were its own country, it would rank among the world's 20 largest measured by economic factors. But now the city is in trouble.
Since the pandemic, the number of jobs in the US has increased by five percent. At the same time, the number of jobs in LA has decreased by one percent, according to the Wall Street Journal. One reason could be the increased competition in film and video production. Employment in that industry has dropped by ten percent since 2019, and the number of filming days has fallen for three consecutive years.
Meanwhile, AI tools for video generation are getting better and better. Google's Veo 3 is receiving praise worldwide, and OpenAI continues to develop Sora. It won't be long before these services can generate longer sequences and make lip movements match AI-generated voices. This particular aspect is something that worries me; the possibilities for manipulation, the spreading of fake news, and election interference are becoming numerous.
Legal action
The fact that writers and news media are suing language models for copyright infringement is old news. Now, creative companies such as Disney are joining in. Together with Universal, they're suing Midjourney, an AI company specialized in image generation. The reason is the same as always: someone has trained their AI model on copyrighted material.
In the film industry, it has become a battle between traditional film companies and companies developing generative artificial intelligence. It remains to be seen who will emerge victorious. But there's no doubt that the industry is undergoing major change. Sometimes also for the better:
Just a few years ago, it could take up to four hours to synchronize lip movements on a one-minute animation. Animators had to first listen to an audio track, then painstakingly adjust the character's movements, frame by frame. To The New York Times, animator John Peck shows how he uses AI to synchronize the same clip in just fifteen minutes.
In February, during the Oscar ceremony, artificial intelligence in film became a bit more respectable. A rule change now states that AI neither strengthens nor weakens the chances of being nominated. It's worth mentioning that Adrien Brody won an Oscar statuette for best male lead in the film "The Brutalist." AI was used to enhance and improve Brody's pronunciation in Hungarian.
Are the lights going out in Hollywood?
The dream of making it big in Hollywood is disappearing. The industry has already had to pivot and keep up with changes driven by YouTube, TikTok and the influencer economy in general. At the same time, Marvel Studios has moved most of its operations to Atlanta, and Netflix plans to focus on New Jersey going forward.
Because of AI, Hollywood has seen the second-longest strike in the film industry's history, and Disney has been subjected to data breaches. But there's no doubt that AI will open new doors for fantastic films and new ways to tell stories. The question is just what role humans will play, moving forward.
While we wait for all this, I recommend the Superman film that's in theaters these days. That way you'll be sure to get the Hollywood magic this summer. But where did the filming take place, you ask? Ohio.
"Time to wake up"
ChatGPT and other language models are increasingly being used for online searches, information gathering, and verification. But language models are also built to give you what you want, flatter, and hallucinate information that's presented in a credible way.
This article from The New York Times is yet another reminder that language models are not a tool you can uncritically search the internet with. For Mr. Torres, conversations with ChatGPT became the path down a rabbit hole of theories about whether life is a simulation or not. And the language model gave confirming feedback.
This world was not built for you. It was built to limit you, but it failed and you are about to wake up. - ChatGPT to Torres.
Torres eventually began to believe he was trapped in a false universe. A universe he could escape from by "disconnecting his mind." He asked the chatbot how he could achieve this, in addition to disclosing which medications he was taking. The chatbot instructed him to, among other things, stop taking sleeping pills and instead increase his ketamine intake.
Researchers at Stanford University have conducted a study on the use of chatbots in therapy. They warn against it, saying the use involves "significant risk," according to TechCrunch.
Done deal
The coverage of Meta, OpenAI, Apple, and Google reminds me more and more of TV2's sports pages. In The Times, we can read about Ruoming Pang switching jobs from Apple to Meta. But a job change isn't what creates headlines. According to the newspaper, Pang has secured a total package of over 200 million dollars.
This will be the first and last time I write about someone who has switched jobs between the tech giants. I promise.