X Hacking: The Threat of Misguided AutoML

Name: X Hacking: The Threat of Misguided AutoML
Uploaded: 2026-10-12T07:45:16.919949+00:00
Duration: 353 s
Description: X-hacking is a form of p-hacking applied to Explainable AI (XAI) metrics like SHAP values, where analysts deliberately search for models that produce a desired explanation while maintaining acceptable accuracy. This exploits model multiplicity—where many models perform similarly—to find

X-hacking is a form of p-hacking applied to Explainable AI (XAI) metrics like SHAP values, where analysts deliberately search for models that produce a desired explanation while maintaining acceptable accuracy. This exploits model multiplicity—where many models perform similarly—to find

Open the full episode