Prioritizing human-centered tech innovation (www.techradar.com)

TechRadar Pro expert argues that the AI problem today is less about algorithms and more about values: up to 85% of AI projects fail not from technical limits but from "value fog"—systems built on idealized user behavior that break trust, morale or real-world usefulness. The author proposes a three-pronged stance: recognize that current AI exposes pervasive values choices (we optimize measurable metrics over meaningful outcomes), treat the AGI race as institutional rather than purely technical (societal readiness and cross‑sector governance matter), and embed multidimensional value frameworks into design so super‑intelligence doesn’t amplify the wrong optimizations. A concrete example: a high‑performing fraud detector that erodes community trust illustrates how behavioral harms negate technical success. For practitioners and policymakers the implications are clear and immediate: move from bolt‑on ethics to values‑first design, build mechanisms for stakeholder participation, and replace command‑and‑control leadership with system‑architects who orchestrate multidimensional value. Technically this demands shifting objective functions beyond single metrics, operationalizing trust and human outcomes as design constraints, and investing in institutional governance and democratic input processes before AGI scales. The bottom line: the future advantage won’t be raw capability but the ability to translate AI into collective human value through inclusive, behaviorally informed design.