AI-recommended features for bias detection

When configuring bias detection in a Predict experiment, you can now use AI to recommend which features to include in your bias scan. The AI analyzes feature names and data samples to identify features that may be relevant for bias analysis, such as those related to age, gender, ethnicity, or other sensitive attributes.

Recommended features are pre-selected for your review. You can adjust the selections before running your experiment, and manual feature selection remains available for bias detection.

This feature uses AI and can make mistakes. Review all recommendations before running your experiment.

Prerequisites: AI Features in Qlik must be enabled by your tenant administrator.

Using generative AI to recommend features for bias detection