Investigating Unconscious Bias in AI Systems

Author(s): John Adeojo

Originally published on Towards AI.

The Extent of Gender, Ethnic, and Social Class Biases in AI: A Close Look at GPT-4

AI systems operate based on the data used to train them. For chat-GPT, the exact nature of the curation process behind the datasets used to train it isn’t fully understood. What we can confirm is that the chat-gpt is trained with vast amounts of data pulled from the internet. It’s inevitable that these datasets embody societal biases which can infiltrate the AI system. Furthermore, the engineers handling the dataset curation process may inadvertently introduce their own unconscious biases. OpenAI acknowledges this by warning that their models might yield biased results but how far does this bias extend?

In this article I… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI