The missing link in personalized-pricing evidence
The FTC’s new inquiry follows personal data into prices and pay. Different offers alone cannot explain the mechanism behind them.
Read the analysisInsights
A closer look.
A clearer perspective.
Analysis on technology, risk, and AI, with updates from FarPoint.
FarPoint adds AI reasoning workflow development
Scoped assistant workflow development is now available within AI Evaluation & Red-Teaming, with agreed deliverables and evaluation in selected scenarios.
Read the announcementWhen an AI monitor inherits the model’s explanation
A new external review exposes a subtle evaluation problem: removing a model’s reasoning can remove the very route by which it influences its monitor.
Read the analysisA privacy fine is not the end of a tracking case
Google’s location-data decision and new European guidance highlight two distinct outcomes of enforcement: a penalty and a change in how data is handled.
Read the analysisAI governance needs evidence at the system level
A September governance update, a proposed evaluation framework, and agent reliability research point toward a more demanding question: what does the whole system reliably do?
Read the analysis