The rise of generative AI (genAI) technologies is reshaping how we perceive and interpret social realities, including modern warfare. While much of the academic focus has been on the military applications of AI, the growing adoption of genAI is also transforming how wars are portrayed, remembered, and understood. Recent scholarly inquiries have raised concerns about the potential of genAI to distort the representation of mass violence, particularly by sanitising and homogenising it. However, little is known about how these representations vary between Western and non-Western genAI models. A new study by Mykola Makhortykh and Miglė Bareikytė explores these questions using the Russian aggression against Ukraine as a case study.
The study audits how two image-generative models—the US-based Midjourney and the Russia-based Kandinsky—represent both fictional and factual episodes of the war. The researchers analysed the models’ responsiveness to war-related prompts, as well as the aesthetic and content-based aspects of the resulting images. Their findings reveal that contextual factors lead to variations in the representation of war, both between different models and within the outputs of the same model. Despite these variations, consistent patterns of representation emerge, which may contribute to the homogenisation of war aesthetics.
The study highlights the potential risks of genAI in distorting the representation of mass violence. By sanitising and homogenising war imagery, these models could create a distorted and sanitised view of conflict, potentially influencing public perception and memory. The researchers emphasise the need for further investigation into how genAI technologies shape the representation of violence and conflict, particularly in the context of ongoing wars like the Russian aggression against Ukraine.
The implications of this research extend beyond academic interest. As genAI technologies become more integrated into media and communication, their influence on public perception and memory of war could have significant geopolitical and social consequences. The study underscores the importance of understanding and addressing the potential biases and distortions in genAI representations of conflict. By doing so, we can ensure that these powerful tools are used responsibly and ethically, preserving the integrity of historical and contemporary narratives of war. Read the original research paper here.