Regenerate Response

To reveal and manifest itself, science needs human guidance for training its models.
Regenerate Response reflects on who is determining the stories and images that are used as data for Artificial Intelligence models, and with what purpose.

With the rapid development of new technologies, it is crucial to examine the social consequences of the bias built into the algorithms and models that inform our increasingly automated world. What is ethical in the social development of these tools and new ways of expressing? This exhibition invites the viewer to co-think critically about the impact of these new forms existing and imagine what alternative visions for the future we really want to create.

The research for this show results from an intersection between the work and literature of Chimamanda Ngozi Adichie, Kate Crawford and Ngũgĩ wa Thiong'o.

The investigation starts with The single-story, the concept that Adichie analyses as a historical way to carry biases and incomplete pre-established formulas that can distort our understanding of the world and perpetuate forms of social inequality. She expressively highlights the danger of a simplistic and one-dimensional narrative that reduces complex individuals and cultures to a narrow set of stereotypes.

Humans are completely impressionable and vulnerable in the face of a story, according to Adichie. In the listener's mind, a story becomes an image, that subconsciously matures to a lived experience. And there is a thin line between what is real and what is imagined.

In his book Decolonising the Mind: The Politics of Language in African Literature, Ngũgĩ wa Thiong'o argues that the colonial project was not just about the physical occupation of land, but also about the colonisation of the mind. And the mind is created largely by images and the emotions attached to them. The collection of the images and information that are fed into the mind determines various levels of the realms of existence. From what our personality can be to the collective unconscious development of an entire nation.

Now, how does the story of global image collection affects AI? In Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence, Crawford argues that Artificial Intelligence is not neutral, but rather reflects and reinforces the values and interests of those who develop and deploy it. In other words: the story repeats itself and builds upon existing power structures and biases, even with a potentially unbiased tool. Crawford highlights the ways in which power dynamics, biases, and inequalities can shape the design, implementation, and use of AI, and how these factors can lead to unintended consequences and extreme negative impacts.

The coding plurality depends on the intention of who is feeding it. Regenerate Response invites Alberto Pereira, Anna Bolta Pascò David McFarlane, Igi Lola Ayedun, Irini Kalaitzidi, Silvana Mendes and Vinicius Couto to imagine, create and compose new datasets, statistics, and images, that tell diverse stories and construct machines with somewhat of ideology. Their research enhances individual’s right and duty to co-create with new technologies.