AI’S NEW FRONTIER: collecting early artefacts of generative artificial intelligence 

In the Design and Digital curatorial team here at the V&A, we’ve been thinking a lot about Artificial Intelligence (AI). For many years, V&A curators have been working to collect and interpret objects and artworks that represent and critically engage with the place of AI technologies in everyday life. For the last two years, we’ve also been watching developments in generative AI – a type of artificial intelligence (AI) capable of creating new text, images, audio, and video content based on users text prompts – following the public release of text-to-image models such as Dall-E, Stable Diffusion and Midjourney, and chatbot ChatGPT, in 2022.

Generative AI applications are built on transformer models, including large language models, that “learn” patterns and context from massive amounts of existing data, including text and images scraped from the web. By identifying patterns and relationships in the data, such as how words usually appear together in a sentence, these models can generate new, statistically likely content in response to users’ text prompts. While AI technologies are not new, for the first time, these mass-market generative AI tools promise free, instant, and limitless content production for general users without the need for programming skills, specialist hardware, or access to large datasets. These developments matter because, as AI access shifted from specialist art, technology and research communities to the general public, changing use patterns catalysed public, media, and political interest in AI, marking a moment of technological transition.

Collecting an object of generative AI

In 2023, we acquired an issue of The Economist magazine from June 2022, which features a cover image generated by an early version of Midjourney.

The cover design of the Asian and American edition of the magazine, and its title, AI’s NEW FRONTIER, introduce articles on AI foundation models: large-scale, general-purpose AI models that signalled breakthroughs in AI development and creative, non-human intelligence. The issue was released shortly before the explosion in generative AI use by a curious public, which led to the widespread sharing and viral circulation of AI-generated images. The prolific interest in potential applications of generative AI ignited debates over the nature of creativity, authorship, copyright, and whose jobs might be replaced first. Against this backdrop – and with The Economist’s editorial remit of global business, economics, current affairs and technology – the magazine’s cover design announces a new era of AI and introduces articles that capture important debates about the future of creative work.

Acquired through the Rapid Response Collecting initiative, the magazine brought together several of the Design and Digital team’s curatorial interests: technology and society, the changing nature of work, and artists’ work with advanced technologies. The V&A collection features several artists using and critically examining machine learning technologies and custom datasets, including Sougwen Chung’s 2017 project Memory, which employed Recurrent Neural Networks and robotics, and Jake Elwes’ Zizi Show, which uses Generative Adversarial Networks (GANs). Artists’ projects like these are vital for deepening our discussions and understanding of the role and implications of AI technologies today. Yet, to understand the growing significance of AI-generated imagery in commercial contexts and everyday visual culture, we must look further.

Why we collect

Curators can acquire objects for one reason, but the process often opens up a host of provocations in contemporary digital collecting. The Economist magazine is no exception. It’s about capturing key moments of technological and cultural transition at a time when these technologies are most visible and tangible, both in their use and prominence in public debate. On The Economist cover, generative AI is hyper-visible. Although it wasn’t the only magazine cover to feature a generated image at the time, the Economist team did so deliberately and provocatively to speculate on this new era of AI. This brings us to another key curatorial challenge: capturing the more intangible elements of digital cultural heritage through objects. We might ask: what’s the “thingness” or material culture of AI today? And how can an object meaningfully tell complex stories of a particular moment? These questions matter today as much as for future audiences, who’ll see objects like The Economist issue in very different technological and cultural contexts.

The enduring meaning of magazines

The rapid developments in AI technologies and their public perception remind us of the need to collect—and quickly, before the context is forgotten. Since magazine covers are inherently immediate and ephemeral by design, they often capture defining moments of cultural significance. The V&A holds many magazines, including the first issue of Time Out London, produced in 1968. This issue features an early digital artwork exhibited in Cybernetic Serendipity: the first major international exhibition devoted to art and new technology, held at the ICA in London in 1968.

The museum also has three versions of the July 2020 issue of British Vogue, also acquired through the Rapid Response Collecting initiative, dedicated to those working on the front line during the COVID-19 pandemic. Featuring portraits of three key workers: a London Underground train driver, a community midwife, and a supermarket worker, the Vogue covers capture individual resilience and document the cultural zeitgeist of those extraordinary times. Magazine covers such as these can distil complex global experiences into resonant images that continue to tell stories of design and society into the future.

The Economist magazine is currently on display in the Design 1900 – Now Galleries at V&A South Kensington. You can also request to see the Vogue magazines in our Prints and Drawings Study Room, and the Time Out issue in the National Art Library.