AI Safety
AI Alignment and Safety in 2035: Two Axes, Four Scenarios, and What Would Falsify Them
Two questions decide how AI alignment and safety practice matures by 2035: whether verification moves beyond behaviour, and whether the field becomes an externally certified discipline or stays self-governed by labs. Four scenarios, each with a falsifier.