Harnessing AI’s creative power: new film explores the tectonic shift in tech

In the dead of night, an engineer at a pioneering tech company slips a journalist into his office. He wants her to interview what he believes to be the first sentient AI.

Several years ago, the premise for Quinn Halleck’s upcoming short film ./sigma_001 would have been classed as science-fiction, or perhaps dystopian. Now, though, the exponential advances in artificial intelligence mean that this scenario could barely even be categorised as ‘speculative fiction’. 

Whilst director Halleck is careful not to name any specific companies or incidents as his principal source of inspiration, the film’s concept integrates seamlessly into the real-life clamour surrounding the tech industry in recent months.

Last July saw the dismissal of Google engineer Blake Lemoine after he voiced his belief that LaMDA, Google’s AI chatbot, was displaying signs of sentience. In March of this year, leading tech figures, including Elon Musk and Apple co-founder Steve Wozniak, signed an open letter calling for a six month pause in development due to the “profound risks” presented by AI with human-competitive intelligence. In May, machine-learning pioneer Geoffrey Hinton resigned from Google, publicly announcing his concerns over the current speed and direction of AI developments. 

John D. Michaels as James, the engineer in ./sigma_001. Photograph: Quinn Halleck

In the trailer for ./sigma_001, the engineer articulates his concerns to the journalist using a particularly powerful analogy: Oppenheimer’s creation of the atomic bomb. He points out that the true extent of the bomb’s destructive power was not known until the Trinity Test in 1945. “What this company has made”, he continues, “is going to make the atomic bomb look like a pea-shooter”. 

As he speaks, a surreal, computer-generated image fills the frame: isometric planes distort around a smaller cube, their flat surfaces beginning to ripple and bubble, threaded with veins of phosphorescent light. AI’s algorithms are being described in visual terms: a creation that evolves rapidly beyond the scope of its maker’s imagination. 

“Ready to meet the future?” Photograph: ./sigma_001, Quinn Halleck Productions

Halleck describes how his fictional AI, Sigma, visually morphs throughout the film, shifting between faces – a young boy, an old grandmother, different ethnicities – depending on the conversation it is having. “It’s finding itself”, he explains, “scraping the Internet to try to figure out the persona it wants to have”.

A twenty-first century Frankenstein’s monster? 

Halleck envisages the film as a “canary in a coal mine”; an examination of the potential consequences of AI innovation, intended to induce reflection and conversation. Despite the clear focus on the weight of responsibility of the creator, though, ./sigma_001 is by no means a doom-mongering attempt to reinforce the bleaker prophecies in current public debate. 

Instead, it explores the existential question of AI’s impact on our lives while simultaneously testing the limits of AI’s creative capacities. It is an instrument employed throughout the film’s production, from initial conceptualisation and story-boarding all the way through to pre-release marketing choices. 

Chase Simone as Sigma, whom the engineer believes is sentient. Photograph: ./sigma_001, Quinn Halleck Productions

The use of AI in cinema has tended to focus around two extremes. At  one end, we have the huge, widely-publicised blockbuster moments, such as the de-ageing of Harrison Ford in the latest instalment of the Indiana Jones franchise using ILM FaceSwap technology.

A de-aged Harrison Ford in Indiana Jones and the Dial of Destiny (2023). Photograph: Lucasfilm Ltd. - © 2022

At the other end, there are the experimental shorts: exploratory and often absurdist, probing the possibilities of AI’s capacity for story-telling. Take, for example, Waymark’s The Frost, the unnerving result of feeding a script into OpenAI’s art-generating AI model DALL-E 2.

Image from The Frost, created using DALL.E 2 AI technology. Photograph: Waymark / Latent Cinema

Halleck’s ./sigma_001 opens up a new position on this spectrum: the use of AI isn’t concentrated into one or two moments of bold visual imagery; nor is it the principal creative driving force. Instead, AI has been employed as a tool across all aspects of the workflow. Some of the technology incorporated into the production includes Chat GPT for prompts, Midjourney for concept art, and DeepFaceLab and Stable Diffusion for visual effects. Halleck also collaborated with voice-generative company Eleven Labs to create a specific sound for Sigma’s speech.

Democratisation of cinema

Halleck sees these technologies as a generous toolkit that can completely transform the capacities of low-budget, independent filmmaking. Ultimately, the film aims to explore AI’s potential to amplify artists and enhance creativity.

At the beginning of his career, Halleck was mentored by director Michael Bay, working with him on films such as 6 Underground and Ambulance. On these huge, high-budget productions, Bay’s resources were impressively extensive, both in terms of crew and technology available.

"What was really special about working on ./sigma_001 was that I was able to replicate that process with almost no money. I suddenly had a concept artist at my fingertips, a writer’s room. All these workflows and pipelines that were traditionally reserved for the elite."

- Quinn Halleck

AI meant that Halleck did not need to compromise his creative vision because of money. In fact, AI enabled Halleck to deepen certain areas of the film’s narrative. One particularly surprising AI contribution was in character development. Halleck asked Chat GPT for ideas on the background story of the journalist, Kat. The language model responded with a detailed suggestion of an event during the character’s time at high-school: specifically, Kat’s investigation into the school authorities’ covering-up of the maltreatment of her best friend. This incident, GPT suggested, could be at the root of Kat’s decision to pursue truth through journalism. Whilst this idea does not appear in the film, it proved a useful prompt for actress Maria-Elena Laas in her preparation for the role of Kat.

The engineer’s reference to Oppenheimer, too, involved experimenting with Chat GPT for ideas during script development:

Part of Chat GPT's response when asked for reasons why sentient AI could be a ‘bigger deal than the first atomic bomb’. Source: Quinn Halleck

To many, the chatbot’s detailed responses will seem eerily self-aware; for Halleck, though, this conversation was another addition to his creative toolset. He picked through the ideas generated by Chat GPT to shape the engineer’s dialogue when attempting to convince Kat of the gravity of Sigma’s potential sentience.

Mitigating the dangers ahead

When envisaging AI’s potential to dramatically broaden the capacities of filmmakers, it is of course impossible to avoid the pressing fears that human roles within the creative industries will be threatened as a result. As demonstrated through the current WGA and SAG-AFTRA strikes in Hollywood, there is a strong concern that AI’s ability to reduce the manual work, time, and (of course) money involved in a film’s production simultaneously means that studios can cut their dependence on human contributions. Even if crew-members’ jobs aren’t immediately stripped away, when the production time of a film is cut from, say, ten months to five months, their source of income becomes increasingly precarious.

SAG-AFTRA and WGA picket line in Los Angeles. Photograph: REUTERS/Mario Anzuoni/File Photo

Halleck is in no denial about these concerns, and he hopes his film will show that great caution is needed across the industry when choosing the next steps to take. Following his numerous discussions with AI pioneers – such as Stability AI’s CEO Emad Mostaque and tech innovators at MIT – Halleck reflects that “this is the biggest thing to happen to us in a long time, and there’s not one single person that knows what’s going on”. He agrees with the call made in March for a six-month pause on AI developments, in order to let legislation and safety-measures catch up with the tech.

Overall, though, Halleck is optimistic. Regarding concerns over job-losses, he views this as a moment of transition, comparable to other rapid technological advancements in recent decades:

"New industries will be created around this new pipeline… There will be job loss in this process, the same way that there was job loss when the Internet was invented. Suddenly you didn’t need as many mail carriers, because there was the email. But what happened? Because of the Internet, so many more jobs were created."

- Quinn Halleck

Indeed, in the 2000s the Internet created 2.6 new jobs for every one destroyed. The current AI transition, then, is crucial to get right. Halleck’s film is testing the waters, placing a spotlight on the balance that needs to be found between harnessing the creative potential of AI and crafting a sustainable future for creative workers. Legislation is needed – and fast – to preserve human contributions and to ensure that worker’s rights and career stability aren’t sacrificed in the name of studio profit. 

./sigma_001 will be released in three parts on DUST, starting from September 12th. 

By Rosa Haworth