The new programme will unite researchers and practitioners around model-explainability and audit methods.
Researchers in alignment and AI law have joined the committee.
Teams presented solutions reducing model-inference energy use by up to 40%.