AI systems amplify unreliable statistics when they appear in training data.
Develop a verification layer that scans for suspicious claims and tracks their provenance.
Academic institutions and responsible AI teams would pay for cleaner data inputs.
MVP could be a browser extension highlighting unsourced stats.
Challenge is building authority in an unregulated space.