Generative-AI, the media industries, and the disappearance of human creative labour

ABSTRACT

This article addresses the transformative role of Generative-AI (Gen-AI) in the creative media and arts industries, focusing on concerns about the disappearance of human creative labour. It critically examines the discourse of the 2023 Writers’ and Actors’ strikes, which replicates prevailing assumptions of the superiority of human creativity over Gen-AI. This discourse emphasises a ‘replacing tasks’ model, anticipating a future where AI assists human creatives in a limited capacity. Against this background, the article applies the ‘meaningful work’ framework to provide an approach to human-AI coexistence which values amplifying human creativity rather than merely supplementing (or supplanting) it. This framework is a conceptual shift which more convincingly recognises and values human contributions in the media industries. Drawing on historical parallels, such as the transition to digital visual effects during the production of Jurassic Park (1992), the article demonstrates how transformational technologies can transcend mere task simplification. The article underscores the importance of creative artists actively finding ways to begin clearly articulating the specific details of human creativity that comprise their artistic agency, and therefore the article advocates an approach that theorises the intrinsic value of human contributions to the media industries while accommodating Gen-AI.

KEYWORDS:

• creative arts
• Gen-AI
• media industries
• work
• employment
• cinema production

Introduction

With the mainstream attention on Generative-AI (Gen-AI) ushered in by the late 2022 release of the chat-interface text generator platform ChatGPT, the creative arts industry’s attention was immediately drawn to what level of disruption this would have to workers. Gen-AI had been producing somewhat convincing text, as well as close to photo-real images, for some time before ChatGPT but the easy-to-use interface and browser-based accessibility made it simple for many people to enter a prompt and marvel at the apparently magical near-immediate response that seemed like genuine human writing (Leaver and Srdarov 2023). Then a push-button AI image-generator churned out a prize-winning artwork, which cued responses of disapproval that the artist had cheated (Roose 2022). Then from May to September of 2023 members of the Writers Guild of America took strike action and, although the purpose of the strike was about more than AI, their stance against studios using ChatGPT to generate scripts brought attention to how creative arts workers feared being ‘replaced’ by AI (Rogin and Corkery 2023). In July, the Screen Actors Guild joined the strike, concerned about the potential for actors to be ‘replaced’ by AI-generated characters modelled on their likenesses (Yeomans et al. 2023).

The incorporation of Gen-AI into the creative arts industries may not be as simple as solving the ethical issue of copyright and reconciling an unpaid debt to the artists’ whose work was used to train the AI models without their consent (DACS 2024), nor by Hollywood studios simply promising not to replace someone with AI and encourage workers to ‘leverage’ the technology to work more quickly if they choose. Certainly, the possibilities, limitations and aesthetic transformations associated with human creatives and AI working together are being explored in relation to writing, professional publishing and visual arts (see Jackson 2017; Jiang et al. 2023; Kendall and Teixeira da Silva 2024). However, it remains for this to be explored in the domain of the creative screen and media industries. This article shows that the problem really focuses on two areas: industrial impact (copyright, jobs, payments) and the threat to perceived meaningfulness of human-created work. These concerns both hinge upon the persistent value of the essential human component of a media product both in the creation as well as its reception. Indeed, the essentialist significance of the human is framed in almost moral terms by Michelle Rae, director of Equity for the Media, Entertainment and Arts Alliance, who states that although AI ‘is absolutely booming and it’s something we can embrace and use in real life […] people deserve to have human stories told by humans’ (Yeomans et al. 2023).

The shock felt by those in the creative arts is understandable given that prior to the time of these events, the consensus in professional creative communities was that their uniquely human complex creative capacities could not be replaced by AI (Bringsjord, Bello, and Ferrucci 2003; Markuckas 2022; Zylinska 2020). Indeed, this attitude was so entrenched that much of the literature attempted to find ways to justify AI as creative, attempting to problematise anthropocentric assumptions of creativity itself (see Mazzone and Elgammal 2019). Now, nobody seriously doubts that Gen-AI will impact the screen industries, particularly with the announcement of OpenAI’s video Gen-AI Sora. Nonetheless, the human worker remains primed as the creative figure or originator in mainstream discussions and opinions. Importantly, the centrality of the human contribution is also believed to be important for the audience’s engagement with the work. According to one professional title designer, audiences ‘try to empathize or connect on a mental level with the origin [of the story, but] When you know that this thing you’re experiencing was not generated by a human, it interrupts that process’ (Landekic, in Johnston 2023). While these opinions are no doubt predominant, they rely upon long-standing Romantic assumptions of authorship (Coeckelbergh 2023) and in doing so repeat vague ideas of human creativity without specifying for contemporary audiences (or, for that matter, funding and finance stakeholders) what this actually is.

A thought-out discussion on this topic is important because the mainstream attitudes of the media industries outlined above arguably cling to an ill-defined yet elitist view of human creativity. As such, there may not be a great deal of sympathy for human workers in creative fields. After all, as AbuMusab (2023) notes: ‘When technology came for the jobs of the working class, other than the people it affected, the resounding chorus among the elite was “retrain, continuing education,” or it is “inevitable”’ (1). One reading of Rae’s comment above is that using ChatGPT to write an email that you will be sick from work one day is an acceptable ‘real life’ use, but it is a snub to audiences to use an image generator to create an animated film in which a character is portrayed with the human emotion of grieving. In this reading, it is somewhat unclear why an everyday person’s email – as a so-called ‘real life’ task – is valued differently to the work of a creative artist. This is especially important for gig workers in the creative fields who already undertake disproportionate amounts of unpaid labour maintaining the ‘relational’ aspects of their work to satisfy clients (Alacovska, Bucher, and Fieseler 2022). Sociological aspects such as these must be considered in relation to Gen-AI’s potential to create the impression that ‘anybody’ can produce creative output, particularly after OpenAI’s Sora video generator announcement in 2024.

As shown below, the creative arts industries have reluctantly embraced the idea of Gen-AI as long as it replaces simple, mundane and repetitive tasks, presuming that this will save creative human individuals from being replaced themselves. However, this approach can be understood as myopic in two ways: first, in terms of how ineffectively it solves the threat of Gen-AI; and second, in the way it misses an important opportunity to enhance human creativity. This article therefore uses a framework of ‘meaningful work’ (Yeoman 2014) to reconcile this conflict between AI and human creativity in the media industries. Meaningful work, a sociological concept which highlights the complex interplay between objective factors (such as job security and fair compensation) and subjective experiences (such as a sense of purpose and pride in one’s work), provides a framework for understanding the challenges and opportunities of human-AI collaboration in creative fields. By reading the dominant response to Gen-AI through the framework of meaningful work, this article argues that the creative arts industries have made a mistake by rushing to embrace what they hope to keep the human in control and keep the AI contributions to replacing minor, mundane and simple tasks. The argument below uses this framework to provide a theoretical justification of the actual significance of media workers’ human creative capacities, moving beyond simply asserting essentialist assumptions of human superiority over the machine. Indeed, there exist prior examples of genuine partnerships between human creatives and digital technologies which show the possibility of producing higher quality creative work by leveraging the creative affordances of both human and machine. After reviewing the current state of Gen-AI and the creative arts industries and summarising the relevance of the meaningful work framework, this article then demonstrates via the case study of visual effects postproduction in Steven Spielberg’s Jurassic Park (1992) how an alternative approach to Gen-AI might develop the possible conditions for genuine partnerships that recognise human contributions to creativity while enhancing them via opportunities also afforded by the technology.

Background & literature review

While much of the public and industrial discussion has focused on issues of ethics (job theft and labour theft), the academic work in this area has primarily sought to establish whether AI can ‘be creative’. The fears and expectations of many creatives are perhaps well summed up by a comment from director James Cameron, famous for the AI dystopias in Terminator (1984) and Terminator 2 (1991) who does not believe Gen-AI could output worthwhile screenplays:

A disembodied mind that’s just regurgitating what other embodied minds have said - about the life they’ve had, about love, about lying, about fear, about mortality […] and just put it all together into a word salad and then regurgitate it […] I don't believe that’s ever going to have something that’s going to move an audience. (Cameron, interviewed in Kapelos 2023)

This points toward what is felt to be the main threat of AI in creative fields: a feeling that Gen-AI is soulless. However, it will become clear that the industrial concerns of theft can be understood as a more proximate issue. To frame the discussion below, this section briefly overviews current perspectives on ethics in relation to Gen-AI and creative work, the assumptions about challenges to human creativity posed by Gen-AI and then frames this against the background of relevant scholarship in meaningful work.

Gen-AI and the ethics of creative labour theft

The creative fields are not so much concerned about extinction-level threats such as AI turning everybody into paperclips (Bostrom 2003), which is somewhat ironic because cinema and literature has explored dystopian fantasies for well over a century. At this stage, the proximate threats to the creative media industries seem to be those areas that make sense to involve a large language model such as ChatGPT or other big data analytics. This includes for example, scriptwriting, idea-generation and analysis of market trends. Meanwhile, there have been some implementations of simple narrow AI-automation for several post-production tasks, such as tagging footage and enabling editors with limited coding ability to automate some parts of their workflow. Netflix uses various AI tools to analyse soundtracks for the purpose of improving captioning as well as industry requirements of identifying music usage etc (Netflix Technology Blog 2023). In the field of visual effects, so-called ‘deepfake’ technology is starting to speed up the time-consuming (but artistically rewarding) process of face replacements for actor’s stunt doubles (Failes 2020). Algorithms can also now adjust the lip movements on an actor’s face to match re-recorded dialogue (either by the same performer, another performer or even in another language) (Helm 2023). These innovations have generally been welcomed by the industry, unlike ChatGPT and the future possibility of AI replications of actors.

However, post-production techniques such as AI lip sync and dubbing point toward the key problem of ‘theft’ in the media industries. For example, if an algorithm can re-voice a video fluently in another language, this is regarded as constituting the theft of a voice performer’s actual work and livelihood. When the Screen Actors members joined the Writers strike in 2023, this threat of job theft, broadly defined, was fundamentally the two Guilds’ major issue with Gen-AI. Actors were concerned that AI-generated digital likenesses may replace them in major roles, a scenario that is literally the plot of Ari Folman’s 2013 film The Congress in which an ‘out of work actor […] sells off the rights to her digital self’ (Broderick, Bender, and McHugh 2018). In addition, the Guild expressed concern over the potential of Gen-AI to replace background actors, often referred to as ‘extras,’ effectively eliminating their roles in the workforce (Rogin and Corkery 2023). The Writers complaints centred on studios intention to use Gen-AI to rewrite scripts, generate basic plot-ideas or eventually create entire scripts which would take away paying jobs of human writers (Merchant 2023).

This industrial perception of job-theft is crystalised by one intriguing instance: when Marvel studios made the bad-taste decision amid the Strike to promote the title sequence of its series Secret Invasion (2023) as being made with AI, they faced immediate backlash including from artists who had worked on earlier parts of the show. The visual effects vendor contracted to make it, New Method Studios, felt the need to issue a public statement that included the following reassuring message: ‘AI is just one tool among the array of tool sets our artists used. No artists’ jobs were replaced by incorporating these new tools; instead, they complemented and assisted our creative teams’ (Giardina 2023 emphasis added).

Ultimately, the outcome of the strike instated ‘regulations’ over the use of Gen-AI content, for example that studios would not use Gen-AI to create or revise literary material, AI would not receive writing credits and studios would disclose if they were providing Gen-AI material to a human writer for further work. These regulations were in addition to establishing ‘guardrails’ against studios using AI likenesses of actors without their consent and compensation (Coyle 2023).

These industrial issues also played out against the wider background of Intellectual Property theft where many artists and writers have raised ethical complaints about the way Gen-AI is ‘trained’ by algorithmically analysing datasets comprised of original human work without permission or compensation (DACS 2024). Interestingly, in the first case by a group of visual artists against this Gen-AI training, the judge dismissed the lawsuit and recommended filing again because of the language used to express their case. Specifically, the judge ruled that the material generated by AI does not infringe copyright, but ‘the AI training process violates their rights’ (Brittain 2023). Thus the wider issues between creative work and AI are not framed purely in labour terms. Repeatedly, the discourse focuses on an assumption that the audience would only accept material that connects a human artist/story-teller with the audience: ‘Audiences aren’t going to want to watch AI-generated shows that lack a human spark for the same reason it’s not interesting to watch two computers play chess’ (Crabtree and Ireland 2023). It is never explicitly stated or identified in these arguments just what is unique or important about human artistic creations (ie., the ‘human spark’). Rather, we are expected to just believe this essentialist position of human superiority. Thus, ultimately the ethical issues of Gen-AI for those in the creative arts at present hinge upon conceptions of the rights of human artists, creators and performers.

Human labour and the AI-Generated image

As of 2024, there is still a great element of unpredictability in the AI generated image, and even more so in the output of generated video. It is possible to influence the output, for example ‘prompt-engineering’ the artistic style, the general features of the subject being created and some broad compositional aspects such as camera height, distance from subject (Korzynski, Mazurek, and Krzypkowska 2023). OpenAI’s Sora appears to produce convincing video, but details remain unclear at this stage about how available it will be to the public or even the industry or to what extent the user can direct its output. For all the talk of ‘whispering’ to the AI (Bozkurt and Sharma 2023), there is still a significant degree of freedom in Gen-AI output.

A full analysis of the differences between AI-generated and traditional media is beyond this article’s scope, and things are changing daily (Hales 2021). Research and theory will need to keep up with this, and for the present discussion it is worth considering only some of the fundamental issues around the so-called human-like attributes of the AI image. To explore the important and meaningful creative labour involved in screen media, consider André Bazin’s (1960) perspective on what distinguishes the moving image from other types of art:

[cinema] offers a different sort of creativity than we find in the traditional arts. The filmmaker works not with pure imaginings but obstinate chunks of actual time and space (Bordwell 2009).

Bazin’s perspective on what makes a photographic or cinematic image distinct is that there was something there to be captured by the camera. Whether that something is built by the filmmaker or found on location is irrelevant. Despite the obvious problems to this perspective posed by digital visual effects, there are no doubt lingering Bazinian assumptions in audience expectations (Prince 2019) and these are exemplified in audience satisfaction with an actor supposedly performing their own stunts, or unpacking the indexical mystery of 300 (Dir., Zack Snyder 2006) as Ayers (2015, 103) shows that audiences wanted to know ‘just how fleshy the bodies of the actors really were’. These concepts also seem to hold even for directors who are strongly associated with high levels of digital manipulation of the cinematic image, often building physical sets where possible to enable actors and crew to ‘feel grounded’ (Prince 2019). Therefore, this practical conception of media image-making suggests that the work of manipulating ‘chunks of actual time and space’ also produces pleasures associated with creative work in the screen media industries, which will be relevant for the discussion of meaningful work below.

Meaningful work in the time of AI

The sociological literature defines the importance of ‘meaningful work’ around the central theme that when a person experiences their work as meaningful this is linked to positive outcomes in their general life (May et al. 2019). Yeoman et al. (2019) argue that meaningful work is a fundamental human need and emphasise subjective factors such as positive experiences at work and the ability to actualise the worker’s self in ‘serving others’ (Lips-Wiersma and Morris 2009). In constituting what makes work meaningful, Martela and Pessi (2018) also highlight the importance of autonomy, self-realisation and broader purpose. Laaser and Karlsson (2022) suggest that autonomy, dignity, and recognition appear as consistent factors in the research and draw attention to the complex interplay between objective and subjective factors associated with what makes work meaningful. These ideas align with the broader philosophical tradition of Dewey [1938] 1988, which contends that work should enable people to express their individual talents and capacities in a way that contributes to the greater social good. For work to be meaningful and satisfying, Dewey maintained, workers needed autonomy, opportunities to develop and apply their intelligence and a sense of participation in a co-operative endeavour that benefits society (Renault 2017). However, Dewey [1932) 1985] was critical of how work is actually experienced by most people in industrial capitalist society; modern forms of economic organisation tend to make work repetitive, stultifying and disconnected from both individual development and social purpose. From this view, Gen-AI is an objective affront to the subjective experience of creative workers: their dignity (we can just have AI write a script faster and better than you), their recognition (we can have an AI actor that replicates your body perform onscreen without paying you) and their autonomy (you have no say over how we will use Gen-AI).

Beyond the sociological definition, meaningful work can also be understood within other philosophical traditions. For example, Marx’s concept of alienation (Marx 1844; Sayers 2003) provides a compelling framework for understanding the fears and concerns of workers in various creative fields in relation to Gen-AI. According to Marx, alienation occurs because workers are estranged from the products of their labour, the process of production and their fellow workers (Marx 1844). Under capitalist social relations, workers are already alienated from the products of their labour because they do not own or control the means of production (Mandel 1970; Sayers 2011). For creative workers, who we have already seen have a deep emotional investment in the content they produce, the prospect of Gen-AI taking over the production of creative content represents a further erosion of their control over their work and a deepening of their sense of alienation. The fear of being replaced by machines, in this context, is not merely a fear of job loss but a fear of losing the ability to express one’s creativity and find meaning in their working lives.

It is understandable then that the introduction of AI poses new challenges and questions about the future of meaningful work. To date, this specific framework has not been applied to the creative arts, however the scholarship of automation in diverse fields such as warehouse stocking, education, military applications, human resource management and policing offers some insights. Bankins and Formosa (2023) suggest there are three broad options for AI integration into the workplace: ‘replacing some tasks, “tending [or managing] the machine” and amplifying human skills’ (725). The authors argue that the first two do not lead to increased feelings of meaningful work, whereas the third has potential because ‘When AI amplifies a worker’s skills it can support them to complete their tasks, undertake more complex tasks, and utilise higher-order thinking and analysis skills’ (735).

These ideas can be mapped onto the creative arts industry. For instance, using ChatGPT to ‘do research’ for a script as the Writers Guild proposes – arguably, an approach that misunderstands that Large Language Models just predict text and do not ‘produce information’ (Bender 2024) – replaces the tasks of a human writer visiting the library, performing an internet search or otherwise reading source material relevant to the historical and geographical context of the narrative they are writing. Presumably, the replaced task is achieved more quickly, enabling the writer to do the more complex work of writing the script. Creative opportunities that exemplify tending the machine would include the Writers Guild fears of studios using LLMs to produce the draft of a script which is then refined by a person. Alternatively, in an even greater fear-inducing scenario, the studio executive simply tends the machine to read over Gen-AI scripts and push various buttons to gently guide the AI toward the final outcome, displacing the human writer as we know them today. It is easy to find examples of proposed, or actual, uses of these two models, however the amplification model is one which has not yet been explored with any rigour. In the discussion below, I draw upon an earlier example from the film Jurassic Park to illustrate what the amplification model might look like in relation to Gen-AI and creativity.

While it may feel like a luxury belief to be concerned about whether AI could replace, displace or otherwise create job insecurity in the creative arts and media industries, it is not. Afterall, as Laaser (2022) notes, the humanities tradition of scholarship emphasises ‘the creation and experience of meaningfulness as a human condition’ (792). It is easy for a creative person to feel nervous about the idea of enjoying the work. In part this may be that there are three common misunderstandings about creative work:

First, that it starts with a brilliant insight (no, ideas emerge after you start working); second, that it’s not something you can learn (no, creativity is based in habits and mindsets that can be practiced); third, that it’s easy and fun (no, it’s hard work that is deliberate and continuous). (Sawyer 2021, 1).

Given that present attitudes towards AI in the workplace tend to emphasise the ‘replacing tasks’ model, it is important to consider what such tasks might be in the creative arts. Or to put it another way, in a ‘solution looking for a problem’ paradigm (Obrenovic and Stolterman 2018) what are the problems that AI is looking to solve? The hard work and dedication inherent to creative processes are demonstrated by artists who engage deeply in problem-solving, which is a core aspect of their craft.

Creative media professionals like the problem solving involved in what they do. The intrinsic ‘work’ of a creative is perhaps best illustrated by a reflection from Australian Writer’s Guild CEO Claire Pullen, who had recently been demonstrated an AI platform that promised to ‘solve the difficult problem’ of figuring out how the narrative threads in a screen series could logically and effectively get separate characters to converge in a later episode. Pullen suggests that ‘the solving of those narrative problems is the creative work, it is part of the craft’ (Pullen 2023). This implies that these aspects of creative labour are not simply mundane busywork; rather they are intrinsically rewarding and necessarily essential components of the creative mindset. Indeed, the pleasure associated with these aspects of the work is part of the self-identity for people in these industries (Deuze and Lewis 2013). Since enjoyment at work is literally one of the key promises of AI for all industries, where workers are promised they will experience enhanced productivity and allowing a greater focus on complex and rewarding activities, it is therefore essential to conceptualise how to best achieve this transition (Furendal and Jebari 2023).

Discussion: beyond mere coexistence with AI

Against this background of hype, fear and anxiety this discussion proposes a position from which the creative fields can demonstrate their enduring human value instead of simply asserting such status. This is not to say, of course, that AI could never replicate or simulate these activities. Nor does the position adhere to Romantic notions of mysterious creative genius. Of course, the concept of ‘creativity’ as the domain of artistic fields is contested and substantial scholarship has problematised assumptions about the seemingly mysterious spontaneous origins of creativity (Hennessey and Amabile 1998), as well suggested there is a creative ‘flow state’ to everyday life (Csikszentmihalyi 1975; Eisenberger and Shanock 2003) including stereotypically ‘non’ creative work (Boden 2004).

Instead, what is specifically at stake here is one of the core ethical aspects of Gen-AI and the creative arts. Creative workers need not be so coy about the idea that they enjoy their work, that it is rewarding and that they feel proud of their contributions to a media artefact which has entertained, informed or otherwise moved an audience (Luckman 2014). Indeed, in conceptualising meaningful work more generally, Yeoman (2014, 235) argues that ‘society ought to be arranged to allow as many people as possible to experience their work as meaningful’. Therefore, rather than simply working with AI assistants to do the same work they already do (but faster or more efficiently), the position advocated below suggests that creative artists can benefit from applying their existing skills, mindsets and conceptual understandings to evolve the aesthetic and engaging aspects of contemporary digital media, while enhancing the possibilities of meaningful work. To illustrate this framework’s potential application in the field of Gen-AI, this section first considers an earlier technological revolution with enormously disruptive results: the transition to Computer Generated Imagery (CGI) that came about through the work on Steven Spielberg’s Jurassic Park (1993). Although this was not the first film to utilise CGI, its production caused a remarkable change in employment and creative practice which at first threatened to follow the replacing tasks model, then the management model before finally stumbling into the amplification model. Through this case-study, the value of meaningful work as a conceptual frame will become clear in order to reconsider how Gen-AI could be adopted by the media industries to amplify and enhance existing human creative skills and achievements.

Jurassic park, rendering traditional visual effects extinct?

The events from the making of Jurassic Park, 30 years ago, have been reported elsewhere (Failes 2018; Gaycken 2015; Prince 2011), however some important aspects have not been unpacked in relation to meaningful work. During pre-production, Industrial Light and Magic (ILM) planned to use Phil Tippett’s stop-motion team to create the dinosaurs Meanwhile, a separate group of ILM staff were experimenting with various types of CGI, and had used them for some instances of the liquid metal T-1000 in Terminator 2: Judgment Day (Dir., Cameron 1991). Tippett’s attitude at the time was: ‘The computer's great for these hallucinogenic things, but good luck doing a living, breathing creature’ (Kasdan 2022). Soon, the CGI team showed a test of a walking dinosaur and it was immediately clear to everybody, including Tippett, that the CGI dinosaurs would be effective for Jurassic Park. Tippett then became extremely unwell, which he attributes to the feeling that ‘everything that I had ever worked for [was in the] trash can,’ until a few weeks later Spielberg invited him to work on the project supervising the animations because the CGI team did not have specific skills that would be needed to give the dinosaurs expressive characteristics (Kasdan 2022). As the film’s director Steven Spielberg recalls the decision:

He [Tippett] was going to go from somebody who had his hands in the craft of changing armatures [on a puppet] to the greatest dinosaur whisperer in the world. (Kasdan 2022)

What Spielberg means here has important implications for present attempts to incorporate Gen-AI into media production at all levels. Tippett and his team of stop-motion animators understood how to simulate expression via animated movements, particularly to create the impression of character thoughts, moods and feelings. In the reflections by various ILM staff in the Oral History of the Dinosaur Input Device (Failes 2018), it is clear that the stop-motion team had vital expertise in the tactile, physical manipulation of models, for example understanding weight distribution, balance natural fluidity of motion.

It was also soon discovered that the best way to incorporate this embodied knowledge into digital form was to have Tippett’s team use a purpose-built physical small-scale model of a dinosaur that contained motion sensors and resembled a pared down version of their traditional puppets. This became known as the ‘Dinosaur Input Device’ (DID) and enabled the animators to work with their hands to adjust movements in a familiar way but with the data transferred directly to the computer animation workstations. The designers of the DID reflect that:

Animators with stop-motion experience are able to start animating immediately, and usually prefer this setup over the traditional setup [because they] don't have to worry about lights, cameras, or other stage impediments [and can] render the resulting animation using computer graphics techniques for more natural motion-blur, textures, and integration with other elements in the scene. (Knep et al. 1995, 308)

Interestingly, Gaycken (2015) demonstrates that the film’s dinosaur movements do not really replicate paleontological hypotheses, and instead are the animators’ impressionistic interpretations of combinations of other known animal movements. For example, the brachiosaur movements were a creative mix of ‘the long strides and grace of the giraffe, but the weight and mass of an elephant’ (Tippett, quoted in Gaycken 2015, 248). Artistically, the outcome of this combination of traditional stop-motion puppetry with CGI is the result of highly expressive performances of the digital characters, including the narrative-enhancing ‘attitude’ of the dinosaurs displayed through their cheekiness and suggested emotions of impatience (Kasdan 2022). This was not possible with CGI alone at the time, and the realistic textures and subtle movements of the dinosaurs’ bodies was not achievable by traditional stop-motion filming techniques.

Thus, the key point is not that the CGI could not replace the puppeteers. Rather, this historical precedent shows that there was an important element of storytelling that could simply not be done by the technology and required the knowledge possessed by the puppeteers. In addition, there are many examples where both traditional and CGI techniques were combined in single shots, showcasing the most effective attributes of each technique. As Prince (2011) notes, ‘The credibility of the film’s effects enabled Steven Spielberg to linger on effects shots that in earlier generations of film would have been much briefer’ (5–6). Therefore, Jurassic Park thus leveraged the human knowledge and capabilities of one team (from the human hand-made paradigm) with those of the computer-based team (representing a form of automation). Using this amplification model of human–computer creativity, each team symbiotically and mutually enhanced the work of the other, while at the same time facilitating opportunities for meaningful work, as expressed in the satisfaction reported by Tippet’s team and the CGI team (Kasdan 2022).

Implications

The sequence of industrial changes outlined in the Jurassic Park example, though focused on visual effects, crystallises many of the key lessons that can be unpacked and extrapolated to understand the challenges and opportunities for meaningful work in the age of Gen-AI and human-AI creativity. First, the oral history of Jurassic Park reveals a distinct recognition of the knowledge and capacities that Tippett and his entire stop-motion puppetry team had, particularly the embodied and intuitive knowledge that could not be written down or necessarily explained to someone else. Second, Tippett sarcastically characterises Spielberg’s decision to shift him from puppet animator to supervising animators as ‘So you never lost your job. It was just [we’ll] kick you upstairs’ (Kasdan 2022). Although this sounds remarkably like Bankins and Formosa’s (2023) conception of managing the machine, what ended up happening was much more aligned with the amplification model. Gaycken (2015) describes this combination of CGI and traditional effects work as ‘interpenetration’ (242).

Despite the predominance of visual effects in contemporary films, there are often significant financial problems due to the exploitation of labour. ILM, with its decades-long support from George Lucas, has remained a major industry vendor like some other stewardship-based visual effects companies, for example Weta Digital via an ongoing connection to its co-founder Peter Jackson. However, other visual effects companies have not been financially sustainable. Two weeks before winning the Academy Award for Best Visual Effects for their work on Life of Pi (Dir., Ang Lee 2013), the company Rhythm and Hues declared bankruptcy having been subject to the ongoing challenges and systemic crises that are a by-product of Hollywood’s approach of demoting the visual effects industry under a system of ‘flexible accumulation’ (Whissel 2023). As Whissel (2023) argues, this could be a result of neoliberalist practices, including a ‘race to the bottom’ in what has classically been known as ‘runaway production’ (Mirrlees 2013) and has been observed in outsourcing post-production to Southeast Asian companies (Windarti 2021). Arguably however, it may also be contingent upon what has been described as the ‘handmade imperative’ (Comiskey 2015, 52). It is, after all, difficult to see what the visual effects artist is doing on the computer, an observation explicitly made by the producers of Jurassic Park during the early testing of the CGI process (Kasdan 2022). Therefore, if the current popular discourse considers Gen-AI to be capable of writing scripts or making moving images, it is important that creative practitioners develop convincing arguments to maintain their value.

Although this article’s tone is generally optimistic, particularly in its demonstration of Jurassic Park’s interpenetrative transition to digital visual effects, the changes of that era also include the challenging labour practices above which persists to this day. These could be avoided in the adoption of Gen-AI. There are lessons to be learned from ‘just transition’ approaches being devised in the energy sectors as they move from fossil fuels to sustainable sources (Newell and Mulvaney 2013). Indeed, new thinking in these fields focuses on ‘just disruption’ whereby the aim is not only to facilitate humane change for displaced workers, but to build resilience through a ‘bounce forwards’ approach where the local community is placed in a better position than before (Morrissey 2023, 289). Therefore, by developing an amplification model of incorporating Gen-AI into the media industries, it may be more likely that funding bodies, distributors and audiences can appreciate what is meant by the human contribution to the creative work.

Meaningful work with Gen-AI: from assistants to co-existence

Currently accepted models of AI’s potential for the creative and screen media industries (as in other industries) typically view the AI an assistant to the human worker, or is at most under the managing control of the human creative. As shown above, this view is clear in the outcomes of the 2023 negotiations between Hollywood studios and both the Writers and Screen Actors Guilds. This has also been influential internationally, with the agreement recommended as guidelines in other countries (Vann-Wall 2023). Indeed, the ‘victory’ has also been regarded as demonstrating a model for other fields grappling with AI outside of the screen media industries (Anguiano and Beckett 2023). Although the outlined issues were pressing in 2023, there are risks associated with the outcome given its implicit acceptance of a task-replacement model embedded in the agreement.

Against this background, the tendency to simply appeal to an assumed value of human-created work as fundamentally ‘better’ than Gen-AI is perhaps not enough. Neither is pledging to ‘regulate’ Gen-AI and retain some influence over how an AI image-scan or virtual performer is used (Crabtree and Ireland 2023). To the credit of everyone involved in the Hollywood-AI negotiations, the particulars of the industrial agreement are due for review in 2026 which is more than enough time for Gen-AI to substantially develop in ways which will likely make these concepts out of date. Therefore, before the 2026 review, workers in creative fields could be even more assertive in their incorporation of Gen-AI, as well as be much more clear about what is actually important and uniquely human about their contributions to creative production. Instead of replacing mundane tasks, the industries could find opportunities to leverage the technology in two ways. First, use it to improve the types of media productions already being created and/or improve the industrial conditions of employees such as leveraging the technology to genuinely improve the diversity of people working in different production roles (O'Brien and Arnold 2024). Second, invent new aesthetic possibilities, a process which has occurred many times in the past with each successive iteration of technology. As demonstrated by Bordwell (1997), there are numerous examples of how the industry has used whatever technological innovation at the time to amplify existing storytelling skills. Thus, Gen-AI can be understood as a further example of these earlier enhancements to creative possibilities.

Importantly, from a sustainability perspective, the present task replacement model runs a significant risk of reducing the need for entry-level skills that usually enable a new worker to gain initial employment in the creative arts. Even if labour conditions were contrived (or ‘democratised’) so that someone could skip that traditional employment entry pathways (for example, as a production assistant), there are nonetheless implications for training, education and meaningful work in an alternative model of co-existence with Gen-AI (Bender 2023; 2024). For instance, Bailey and Madden (2017) suggest that while the novice tasks associated with beginning a particular work role may seem meaningless at the time, they often mark an important developmental phase, enabling a sense of achievement and progression in addition to mastering the basic skills. As demonstrated above in relation to the industry expectations of Gen-AI, given that the dominant paradigm relies upon a loose assumption that the human is important, this must therefore also involve opportunities for self-actualisation as well as clear pathways from novice to expert. In addition, as Gen-AI matures there may be an ever-expanding divide between professional artists who can pay for the premium platforms and those emerging artists who cannot yet afford the cost.

Conclusion and future research

What is at stake in the integration of Gen-AI into the creative arts industries is that instead of relying on a Romantic notion of the essential value of human creativity, which is often expressed in vague terms, creative artists could more directly (and more specifically) identify the actual contributions of humans to the creative works they produce. For example, the important aspect of the Jurassic Park anecdote is that the creative contributions of Tippett’s team were revealed not to be that they had the skills to manipulate a puppet one frame at a time to suggest the illusion of smooth motion. Rather, their specifically human creativity turned out to be the intuitive and embodied understanding of how to convey personality and emotion through animation, particularly in ways which would be cognitively and affectively understood by human viewers. Understanding and articulating exactly these kinds of things that constitute human creativity will enable audiences and other stakeholders to appreciate why the arts matter. In addition, it will provide clearer directions for future collaboration that takes advantage of the unique capabilities of humans and AI. To build upon ideas from productivity consultant Cal Newport (2024) about AI in the workplace more generally, it is perhaps too simplistic to think of how AI could help write emails faster; instead, we should be considering the ways in which AI could create the conditions of work so that email is simply not necessary.

This article has focused on the creative screen media industries, but similar debates will occur in relation to other media and arts industries (music, dance etc) which may draw similar or divergent findings about meaningful work. Future research should also explore the particular type of fun that AI image tinkerers enjoy as they play with Gen-AI and disseminate their results, workflows and experiences on outlets such as YouTube. In addition, there are many adjacent online creative industries outside the scope of this article, such as the economies comprising influencers on TikTok and Instagram (Gurrieri, Drenten, and Abidin 2023). Clearly, these creative economies rely upon audiences engaging with the medium via a kind of communicative-alliance between performer and viewer/customer (Sætrem 2023). Thus, the ‘human-made’ aspect of such work is essential. A quick search on YouTube reveals numerous ‘how-to’ videos for interested and motivated people to use Gen-AI to create an AI influencer. This was once a multi-million dollar effort by an international marketing company to create Miquela (Blanton and Carbajal 2019; Robinson 2020), but now it is possible to import footage of the body of an existing human influencer’s video and use it to generate a ‘new’ video. Research could explore, for instance, whether it infringes copyright to use someone’s uploaded body movements for an AI-generated motion capture which then drives the movements of an animated character in a monetised video.

This article has argued that framing the issue of Gen-AI primarily as an ethical matter of labour fairness and wage theft misses the opportunity to enhance the future of the creative arts in meaningful terms that amplify the human elements and components of creative work. In addition, centring the discussion on this kind of defensive framework leaves open the risk for the issue to be resolved by Universal Basic Income (UBI). At present the UBI debate is split between two broad positions. One considers UBI to be a solution to previous inequalities intrinsic to the creative fields where newcomers either need existing financial support or take low-paying side work while trying to get their start in the industry (White 2019). From this view, the ‘explosion of creativity’ (Teer 2020) that would result from UBI capitalises on the present community of creative dabblers who pursue expressive outlets as forms of ‘serious leisure’ (Stebbins 2020). Alternatively, UBI has been critiqued as an economic system which provides a basic subsistence to enable the continuation of unpaid digital labour (Mathers 2020). In addition, currently proposed implementations of UBI present difficulties that are unique to particular countries. For example, the US debate seems to be primarily focused on the practicalities associated with the cost of running UBI (Hoynes and Rothstein 2019) whereas a more significant challenge in the UK context is resolving whether or not the concept of UBI is important enough to justify the cost (Reed et al. 2023). These debates are compounded by sociocultural differences (Marais 2022) that impact how UBI’s interaction with existing welfare systems is viewed; some contexts imagine it to supplement existing systems and others see UBI as a total replacement. Certainly, future research will need to explore the possibilities and challenges of UBI in relation to creative fields, and particularly in relation to Gen-AI, given that existing debates frequently ignore the ‘material basis of workers’ anxiety’ (Kelly 2023, 839). Thus, applying the meaningful framework advanced in this discussion will add a valuable contribution to future research and policy directions in UBI.

The amplification model proposed here contributes to debates in relation to Gen-AI by potentially allaying the need for the industry to continually restate and reclaim their intrinsic ‘human’ value, even as neoliberal economic practices continually seek to find cheaper and more profitable methods (including AI) to package consumer culture (Stocchetti 2023). Rather than begrudgingly tolerate Gen-AI as a means to ‘replace some tasks,’ there are opportunities to leverage Gen-AI to enhance the key skills and embodied knowledge that workers in the creative fields possess, at the same time producing better creative and cultural products that otherwise could not have existed. A meaningful work paradigm for Gen-AI does not just assign digital tasks; it promotes an interpenetrative dynamic in continuity with some specific previous technological transformations where the technology supports and enriches human creative efforts, as demonstrated in the case of the visual effects for Jurassic Park. This kind of collaborative interplay aims to bolster human creative output while preserving the core of what makes creative work rewarding: the joy and confidence derived from engaging deeply with one’s craft, supported by the reassurance that such engagement is both valued and essential. Indeed, without this there is the risk that creative work could indeed become meaningless work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Stuart Bender

Stuart Marshall Bender is Associate Professor in Screen Arts at Curtin University. He is a filmmaker with expertise in digital visual effects and teaches screen production and theory in the context of a creative arts degree. Stuart’s research specialty is understanding the impact of high-emotion media content on audiences, and his creative works have been screened both competitively and by invitation at international film festivals.