Amazon, Microsoft lead efforts to tackle inherent biases in GenAI

Generative artificial intelligence (AI) chatbots are not free of inherent biases, and several ongoing projects from the likes of Amazon and Microsoft are working towards reducing these biases linked to large language models (LLMs).

Rajeev Rastogi, vice president, machine learning at Amazon, is leading such projects where several Amazon India employees are working.

The international non-profit Bill and Melinda Gates Foundation is sponsoring a three-year project by Bengaluru-based artificial intelligence startup Karya, which is working on coming up with a method to say with a certain degree of certainty if a dataset is gender-biased or not, and have a pipeline to de-bias the datasets.