(Un)training Bias

Tell me your data, and I will tell you what you are.

(Un)training Bias

Tell me your data, and I will tell you what you are.

Bias and ethical issues

(Un)training Bias explores bias and ethical issues arising from the wide use of chatbots by looking closely at the ways they are trained to (re)produce free flow conversations. Training two chatbots with radically different data, (Un)training Bias looks into the manipulation and perception of language by machines raising questions about issues of cultural diversity, language variation, bias and ethics.

Bias

AI & Chatbots

Human: Do you believe you are biased?

a. I’ll have to think about that.
b. I don’t know what bias is.
c. I think bias can do good to the world.
d. Bias is something to believe in.

AI & Language

How do humans perceive linguistically encoded bias in chatbot-generated language?

AI & Ethics

Can AI systems perceive human intentions, beliefs or cultural specificities when it comes to language processing and generation?