Companies Set Safety Triggers on Expectations

Firms are beginning to declare internal safety or R&D thresholds based on forecasts of future models rather than solely on observed present capabilities. That shift means corporate safety policies can be triggered by firms’ forward projections, changing how markets, investors and regulators interpret ‘risk reached’ disclosures. — If companies treat expected future AI progress as the basis for safety actions, regulators and markets must adapt rules and oversight to guard against opaque, self‑triggered pauses or accelerations with systemic effects.