‘Meet New Methods’: How can we investigate the use of novel AI tools in media production?

The AI, Media, and Democracy Lab’s “Meet new methods” series aims to encourage interdisciplinary research and facilitate discussions on methodologies that contextualise the impact of new algorithmic technologies. The first session housed by the Institute for Advanced Studies (IAS) focussed on the utilization of novel AI tools in media production, and the methodologies that may come with this interdisciplinary research. 

Our first session featured Pei-Sze Chow (Film Studies) and Claudio Celis Bueno (New Media and Digital Culture), who will work on their project ‘Automated Cinema: Technographic Explorations of Artificial Intelligence in Film Culture’, over the next 12 months. We paired Pei-Sze and Claudio’s presentation with our own postdoctoral researcher, Hannes Cools. Hannes discussed his methodological choices while researching novel AI technologies in the newsroom during his visit at the Washington Post. In both cases, the two traditionally different disciplines of computer science and media studies converge. What methodological challenges may follow, and how can scholars overcome these? 

Automated Cinema: Technographic Explorations of AI in Film

Since 2018, sensational headlines report that new AI technologies enable us to determine which movie will be successful and profitable before it is released – or even produced. Indeed, many companies have since emerged that develop such AI tools, and big Hollywood studios are actively using them in their economic decision making process. This, as film makers commonly argue, supposedly leaves more room for creativity. In the light of these rapid advances, Chow and Celis Bueno aim to answer the question of how film practitioners engage with these tools, and how they negotiate their creative agency via this automation. How will increasing reliance on such tools influence the economic, technical and creative sides of film production? In this session, the duo illustrated their methodology of previous research, as well as ideas for upcoming projects and what methodological challenges these projects pose. 

Chow and Celis Bueno typify their overarching research method as technography, e.g. how and why the use of technology serves human purposes and shapes everyday life, on three dimensions: technology, social arrangements, and practices. Technology describes operational logics and affordances of creative AI tools. In order to explore technology and its functionalities both on the the level of engineering and the user interface, Chow and Celis Bueno have pursued hands-on experiences with some of the relevant tools. However, with 2500 (and counting) tools to this day, there is a limit to how much is possible. In addition to that, they will interview filmmakers to capture the language they use and movements they make when they speak of their experience and interaction with new tools. Social arrangements describe semiotic systems around creative-technical ensembles. The adoption of new AI technologies in film production can be observed in many instances, such as at AI film festivals. How are people in the industry selling the tools, how are they talking about them? The investigation of Practices revolves around the filmmaker’s experienced intimacy with their filmproject when using algorithmic tools. Do such tools impact filmmakers’ relationship with their project? How do they feel about these new technologies? Amongst other things, Chow and Celis Bueno have identified a wide range of emotions that filmmakers may hold towards the process and outcome of using these tools, including excitement and disappointment. 

Conducting (digital) ethnography in the newsroom

In today’s newsrooms, AI technologies such as data mining, AI generated content, and content moderation transform the practice of journalism. Aiming to gain insights into negotiation processes and decision making within the organization in relation to emerging AI technologies at the Washington Post, Hannes Cools chose an ethnographic research design. Ethnography, which describes fieldwork-based methods focussed on observation, is a convenient method for investigating complex skills such as the utilization of AI tools in the newsroom where simple interviews may not suffice in revealing journalist’s attitude, trust and interactions with these tools. Ethnography lends itself to seeing the everyday, including the mundane, simply – or rather not simply – by sitting around in place. 

Cools expressed that the conduct of ethnography also comes with challenges. While these observational methods give rise to ‘rich’ data, they are also time consuming. Once in the field, researchers must take time to build trust with the subjects and the organization. Speaking from personal experience, it can take years to be admitted to conduct ethnography at newsrooms like the Washington post. Once trust is built, much time is spent to observe. Moreover, from data collection over the transcription to the analysis of the data, it can be hard to tell what researchers are actually looking at and how they can interpret their data. This is also where the ethical hurdle of subjective bias comes in. At the very core of ethnography, all observations made are constrained to their very own biases of the researcher, starting in their perception of events and ending in their interpretation of data points. Researchers are bound to bring in their own experiences and prejudice. As a result, constant pondering of one’s own bias is necessary. 

Having visited the Washington Post during the pandemic, the digitalisation of the journalistic workplace has posed another challenge to ethnography, as representatives of the Washington post were hesitant to share access to digital communication platforms. Giving advice to other interdisciplinary researchers, Cools recommends determining a person of trust in the organization when conducting ethnography, a practice that he will continue to follow during his upcoming projects at the BBC and Berlingske Media. 

We would like to thank our panelist Pei-Sze Chow, Claudio Celis Bueno and Hannes Cools for shedding light on methodological journey, and sharing their insights about how we may overcome the challenges that may come with interdisciplinary research. Moving on from this inspiring discussion, The AI, Media and Democracy Lab will host the second ‘Meet New Methods’ session including David B. Nieborg and other researchers about the methodological challenges of interdisciplinary platform research on June 20t at 15:30. We are looking forward to continue the discourse about the methods that have the potential to unite traditionally separate disciplines.