by Ella Wright, University of Birmingham
Artificial Intelligence (A.I) is booming. Techniques and technologies are maturing, advancing, and becoming more accessible to everyday public consumers at a startling rate. Take for example ChatGTP, a chatbot made by company OpenAI, which ‘interacts’ with its user in a conversational fashion, and is able to generate, among many things: computer code, film scripts, and improve grammar, all in a matter of seconds. To quote The Economist, “the transition into a world filled with computer programs capable of human levels of conversation and language comprehension and superhuman powers of data assimilation and pattern recognition has just begun”. For the field of digital humanities, by which its very definition is the incorporation of computer technology into research within the humanities, A.I. is quickly becoming an unavoidable aspect of the discipline. Whilst this new tech provides opportunities to create new forms of research that were never before thought possible, it also raises new questions and issues that must be addressed.
As a PhD student whose thesis is researching videographic research-by-practice, a branch of digital humanities perhaps more commonly known as the video essay, A.I. has recently been extremely relevant to my work. The use of an artificial intelligence system to animate photographs was a key component of a recent video essay project of mine titled 'Life and Death of the Image.' The video was inspired by Laura Mulvey’s work Death 24x a Second (2006), in which Mulvey motions that the “stillness of death” is present in cinema, due to the movement of film being an illusion produced by the rapid forward motion of still photographic frames (Mulvey, 2006: 88). From this, Mulvey highlights the paradox of these moving images: there exists an immediacy and advancing progression of the cinematic film played at 24 frames per second, which is parallel to the pastness of each singular static frame. Thus, there is a simultaneous life and death associated with the moving image. Upon reading Mulvey's work, the focus turned to death and photographs. The aim was to identify the most haunting photographs, those which embody the "presence of death" in their photographic stillness (Mulvey, 2006: 102). The registration photographs of the Auschwitz-Birkenau concentration camp were ultimately selected as the answer to this inquiry.
Held by the Auschwitz-Birkenau Memorial and Museum, there exists a collection of 38,916 registration photographs, taken between 1941-1945. There surely can be no greater spectre of death than the one which hangs over these images. However, looking at these images in the context of Mulvey, a question was raised: what would it look like if these photographs moved? Would the motion of cinema somehow add life to these images so tainted with death? So, to answer this question, the photographs obviously had to move. How to do this? A.I. technology. The technology used in this particular case is called Deep Nostalgia, licensed by MyHeritage, a genealogical site for discovering family trees. To quote directly from their website:
“The remarkable technology for animating photos was licensed by MyHeritage from D-ID, a company specializing in video reenactment using deep learning. MyHeritage integrated this technology to animate the faces in historical photos and create high-quality, realistic video footage. The Deep Nostalgia™ feature uses several drivers prepared by MyHeritage. Each driver is a video consisting of a fixed sequence of movements and gestures. Deep Nostalgia™ can very accurately apply the drivers to a face in your still photo, creating a short video that you can share with your friends and family. The driver guides the movements in the animation so you can see your ancestors smile, blink, and turn their heads. This really brings your photos to life!”
The result is, as the website states, a “technological simulation of how the person in your photo would have moved and looked if they were captured on video”. The application is both free and publicly available at the time of writing, requiring only an email sign up to the website to receive access. The utilising of this A.I. technology in this project sparked a range of discussions and controversies, which are of great importance in the ongoing conversations surrounding the implications of using these technologies in the digital humanities (as well as wider) research. The first issue is that of consent. The tagline of the MyHeritage website is “animate your family photos”. But there are no restrictions or security measures to ensure that you know the person in the photograph you are uploading, or that they have given explicit consent for their images to be used. If one uploads a photograph of a person they do not personally know, that individual has not consented to their image being uploaded to the MyHeritage servers, nor has consented to the animation drivers being applied to their faces. Indeed, it would appear that MyHeritage itself potentially encourages the use of this technology beyond your own family photographs, with a part of their advertising featuring an animated photo of Abraham Lincoln. They even took it one step further - colourising Lincoln and making him appear to speak. Did Lincoln consent to his likeness being used to advertise a genealogical website? The general consensus would be that he did not.
In terms of ‘Life and Death of the Image’ in particular, the issue of consent is extremely pertinent. It is, obviously, impossible for the individuals photographed, who are since deceased, to consent to their likenesses being used in this manner. Yet unlike Abraham Lincoln, there are the added implications of these photographs being taken at what was likely an incredibly traumatic instance in these people’s lives. There is also the question of whether this may potentially deprive them further of their agency that was already stripped by the Third Reich regime. Additionally, they also cannot consent to their images being used in my research. The Deep Nostalgia machine also adds qualities such as smiles and joyful expressions to the image in motion which may be critiqued as vulgar regarding the context in which the images were taken. So, from this, we can see the ethical implications not only of the A.I. technology itself, but how these issues can be exacerbated depending on the type of image you enter into the machine.
For ‘Life and Death of the Image’, the counter argument is that this video project has the potential to return a degree of dignity to the individuals represented in the photographs. During the time the photographs were taken, the individuals were stripped of their humanity and were reduced solely to the qualities deemed asocial by the Nazi regime. The use of animation technology may provide some restoration to their personhood. It is not an attempt to impose a smile on their faces with heavy-handed A.I. technology, but rather an effort to give back the smile they may have had before the horrific events took place. Although some individuals may not agree with this sentiment, it does remain a crucial aspect of the project's potential impact.
Other issues remain, however, particularly that of copyright and surveillance. “Existing A.I. systems raise real concerns about bias, privacy and intellectual-property rights.” (The Economist). This is a point of contention in itself, but when considered in parallel with ‘Life and Death of the Image’, perhaps takes on an even more concerning facet. One of the reasons behind the Third Reich regime taking the registration photographs at Auschwitz-Birkenau was that of surveillance, in terms of marking and identifying the prisoners. (see here for more information). There are several points to consider. One, that A.I. algorithms can result in discriminatory practices towards certain groups of people. If algorithms are trained on biased data or rely on pre-existing stereotypes, disproportionate targeting or misidentification of individuals based on factors such as race, ethnicity, religion or gender may happen. A.I. systems can also monitor and collect personal data that is freely uploaded to its systems by the users. MyHeritage states that the photos uploaded are protected and not shared with any third party, although the terms and conditions have not been fully analysed. But what if they were? If the photographs are stored by the company, they have a large database of people’s faces, potentially useful for a myriad of surveillance reasons. There is a potential market for the collection and storage of data through A.I powered systems to be misused or abused by any powerful agents; the uploading of photographs to an A.I. system has feasible links to observational control, whether that be by companies, markets, or governments. When considered in tandem with ‘Life and Death of the Image’, and the surveillance of the Third Reich during the Second World War, the implications are, whilst perhaps leading into the realm of conspiratorial thinking, are nonetheless quite chilling.
Ultimately, the points raised in this article are only scratching the very surface of implications and concerns surrounding the use of A.I. technology, both in general and for the purposes of digital humanities research. The hope is that by continually raising them, that the conversations continue to be had about how these new mechanics can both help and hinder our society and our research.
About the Author
Ella Wright is an artist, filmmaker, and audiovisual researcher based in Birmingham, U.K. They are currently studying towards a practice based audiovisual PhD at the University of Birmingham. Their thesis working title is ‘Half-essay, half-film: the search for concreteness in the video essay form’. www.ellavictoriawright.com
Bibliography
The Economist. (n.d.). How AI could change computing, culture and the course of history. [online] Available at: https://www.economist.com/essay/2023/04/20/how-ai-could-change-computing-culture-and-the-course-of-history [Accessed 24 Apr. 2023].
MyHeritage. (n.d.). MyHeritage Deep NostalgiaTM, deep learning technology to animate the faces in still family photos. [online] Available at: https://www.myheritage.jp/deep-nostalgia [Accessed 24 Apr. 2023]. Mulvey, L. (2015). Death 24x a second stillness and the moving image. London Reaktion Books.
MyHeritage (2021). Abraham Lincoln Discovers His Family History on MyHeritage. YouTube. Available at: https://www.youtube.com/watch?v=kEtiajHLmQY [Accessed 24 Apr 2023].
The Economist. (n.d.). How to worry wisely about artificial intelligence. [online] Available at: https://www.economist.com/leaders/2023/04/20/how-to-worry-wisely-about-artificial-intelligence. [Accessed 24 Apr. 2023].
Faces of Auschwitz. (n.d.). Registration Pictures and Marking System. [online] Available at: https://facesofauschwitz.com/registration-pictures/). [Accessed 24 Apr. 2023].