Policy intervention 1: Increase transparency around the data used to train AI models
Summary
The ODI argues for increased transparency regarding AI training and fine-tuning data to mitigate bias, ensure legal compliance, and foster trust in AI systems.
Key quotes
Deploying models in the public sphere without oversight is negligence too
the user should know when to trust the system’s predictions and when to apply their own judgement
The post outlines the need for standardized reporting of training datasets, including size, source, and licensing. It examines the regulatory landscape in the UK, US, EU, and Japan, suggesting that the UK government adopt transparency tools like Dataset Cards.