This month Solana Moiraghi shared our key findings from a usability test of Meta´s AI integrated on WhatsApp.
What biases become visible when the elderly from a low-income neighbourhood in Argentina interact with a Generative AI model developed in the Global North?
At @socialesuba we presented concrete examples of sociocultural biases in the model’s responses—rooted in the Anglocentric training data—and highlighted how the elderly often show an authority bias, valuing their interactions positively even when the AI’s answers do not fit their cultural realities.
A contribution that invites us to rethink the present and future of AI from a Latin American perspective. Read more here
