Postdoctoral and SHAPE-researcher Nicolas Maleve is organizing a series of workshops aimed at promoting a more democratic and participatory approach to AI, particularly in the realm of image generation. With a background in both traditional art and computer vision, Maleve finds himself at the intersection of artistic practice and technology. This convergence has inspired his latest research and workshop series, which focuses on the decentralization of AI.

Decentralizing AI

Maleve's workshops respond to the growing centralization of AI technologies, which have long been dominated by large tech companies. These companies control the infrastructure and algorithms behind AI systems, raising concerns about monopolistic practices and a lack of transparency. However, as communities of developers, enthusiasts, and activists challenge this status quo, alternative decentralized tools are emerging.

Maleve’s primary question in his ongoing research within the Knowledge Servers project is: How far can we separate AI from the platform economy, and what consequences might this have for a more democratic and participatory culture?

In these workshops, Maleve and his participants delve into the world of AI image generation. The workshops consist of three core components:

1. Technical exploration: The first component is experimental and hands-on. Participants engage with AI image generation software that can be used locally, disconnected from centralized platforms.

2. Understanding technology attachments: The second component focuses on understanding the broader ecosystem surrounding AI technologies. Participants create timelines, diagrams, and map out the actors involved in the development and usage of these technologies.

3. Decentralized governance: The third part addresses governance issues. How does decentralization empower individuals? What challenges arise in taking responsibility for the infrastructure? What difficulties emerge when attempting to decentralize technology?

These workshops involve diverse groups, including students from AU's Alternative Data Futures and Digital Living courses, hacker collectives like Code&Share, research groups such as the Centre for Aesthetics of AI Images, and artists from the Jutland Art Academy.

Overcoming challenges

Designing these workshops has not come without its challenges. One of the biggest hurdles Maleve faces is convincing participants of the importance of these questions.

“Why bother with alternative technologies when platforms give you easy access to polished products?” Maleve reflects. Recent developments have made the need for decentralization increasingly pressing. The growing ties between Silicon Valley and government administrations highlight the importance of understanding who controls AI and how it operates.

Maleve raises a crucial issue: representation. “Why does an image of a street in the US appear with more detail and variation than a street in Madagascar? Why are stereotypes about gender and race so persistent in AI-generated images?” he asks. The answer often lies in the data sets used to train these systems. With access to the right data and infrastructure, these representations can be altered—and decentralizing AI is a step toward that change.

A shift in perspective

Ultimately, Maleve hopes the workshops will shift participants’ perspectives on AI. The aim is not just to spark curiosity but to ignite enthusiasm for becoming active agents in shaping how AI systems are designed and used. “There is no fate that dictates how AI should work,” Maleve states. “We have the power to experiment, question, and influence these systems.”

By the end of the workshops, Maleve hopes participants will leave with a deeper understanding of the possibilities of decentralized AI and feel empowered to continue exploring these tools beyond the scope of the sessions.

As AI continues to play an ever-increasing role in our lives, Maleve’s workshops offer a glimpse into a future where technology is not solely dictated by corporate giants but is instead shaped by a collective effort to build more inclusive and diverse systems.