Tutor/s
Neighborhood People
Neighborhood People
Master Project
- MAIAD 26
In recent years, Artificial Intelligence systems have transitioned from mere computational tools to veritable arbiters of social identity, presiding over crucial decisions in sensitive sectors such as corporate recruitment, access to credit, and public safety. This transition is often accompanied by a strong technocratic narrative that mistakes numerical quantification for a guarantee of absolute fairness. But in reality, systems tend to reproduce and even amplify the stereotypes already inherent in the datasets on which the AI is trained.
Leveraging the synergy between generative models (Midjourney), multimodal classification models and neural networks (OpenAI CLIP), and node-based visual programming environments (TouchDesigner), the project aims to show how Artificial Intelligence systems translate identity and ethnic fluidity into rigid, stereotyped clusters. Trained on images taken from social media, then called upon to classify generic portraits of people from Barcelona and assign them to three characteristic neighborhoods, the AI cascades common biases, homogenizing every nuance of identity.
author/s
