Short Description

SPAM2GAN (2021)  is an artwork by Frederik De Wilde that creatively uses spam emails, processed through Generative Adversarial Networks (GANs) to generate images, combined with text-to-speech technology for an audiovisual experience. This piece challenges viewers to reconsider spam as more than a nuisance, exploring its role in digital narratives.

Artistic Process

De Wilde collects spam from his inbox, using the text to create images via GANs—a type of machine learning for synthetic data. These images are then merged with the original text using Touchdesigner software, and computer voices read the messages aloud, giving voice to the creators behind spam.

Cultural and Societal Reflection

The artwork invites reflection on spam’s evolution over 20 years and its viral spread through social networks, mirroring societal changes. It highlights the human stories behind spam, from entrepreneurs to bots, and prompts questions about data—whether mined like resources or produced by us—and its impact on our digital lives.

SPAM2GAN is as a pioneering audiovisual real-time artwork by Belgian artist Frederik De Wilde, known for his interdisciplinary work at the intersection of art, science, and technology. The piece ingeniously merges the pervasive issue of spam emails with the advanced capabilities of Generative Adversarial Networks (GANs), a machine learning approach adept at producing realistic synthetic data. De Wilde draws from his personal inbox, transforming the text of spam messages into visual imagery via GANs, and enhances this with text-to-speech technology to give auditory expression to these typically silent digital intrusions. This innovative synthesis aims to cast light on another facet of our digital culture’s shadow history, often disregarded but integral to our online experiences.

Spam is frequently regarded as a negative emblem of digital culture, comprising unsolicited messages that range from requests for personal information to scams and advertisements for ambiguous products and services. Since the inception of commercial email in the 1990s, it has remained an enduring irritant, despite persistent attempts to combat it with detection and filtering software. However, De Wilde challenges this conventional viewpoint by proposing that spam could be seen as a form of entertainment or a source of information that contributes to the ever-evolving narrative of our times, offering insights into how digital culture is shaped. His fascination lies in observing how the format and content of spam have evolved over the past two decades and its implications for society at large, noting that spam spreads through social networks like a virus, reflecting erratic changes in society.

For SPAM2GAN, De Wilde has compiled a collection of spam messages gathered over time. The textual content from these messages is used to generate images through GANs, which are then combined with the original text using a custom software patch in Touchdesigner, a visual development platform. Furthermore, computer-generated voices read aloud the selected messages, thereby giving voice to the producers of spam—whether they operate from nondescript office buildings or homes worldwide, encompassing a spectrum from enterprising individuals to automated bots, pharmaceutical swindlers, unethical marketers, and identity thieves. This comprehensive approach not only revives otherwise ignored or deleted messages but also creates an immersive experience that revitalizes these digital communications.

The artwork invites viewers to consider the human narratives behind these digital intrusions and reflect on their broader societal implications. By transforming spam into an artistic medium, De Wilde highlights the complex and often overlooked facets of our digital existence. It stimulates reflection on the nature of data in our contemporary society, raising questions about whether data are resources to be extracted like minerals or are products of our own creation. The artwork underscores our ambiguous relationship with data: how we collect it, how it influences us when we model it through algorithms like GANs, and how it, in turn, shapes our understanding and behavior. This multifaceted approach encourages a reevaluation of our relationship with the digital world, finding beauty and meaning in unforeseen places.