Dr. Kumud R. Jha · Singapore · Doctorate in AI · US Patent Holder View LinkedIn Profile
The Fear vs. the Reality of AI Governance
Day 5 I told you the fear of a more powerful force dominating us is ancient, wired in, and rational. Today I want to answer the question that should follow.
Today I want to answer the question that should follow: so who is watching the watchers?
Because here is what the fearmongers won’t tell you — and what the governance experts are too boring to package well enough for anyone to share.
There are people whose entire job is to make sure AI doesn’t run away from human control. And they are, quietly, making progress.
The EU AI Act is the world’s first comprehensive legal framework for AI — classifying systems by risk level, banning certain uses outright, and requiring human oversight for high-stakes decisions. It passed. It is law. It is being enforced.
The US Executive Order on AI (2023) directed federal agencies to develop safety standards, watermarking requirements for AI-generated content, and civil rights guidance for automated systems. Not perfect. But it existed and moved the needle.
The UK’s AI Safety Institute — the world’s first government body dedicated solely to testing frontier AI models before public release — has already run evaluations on models from the major labs. Including the ones you use every day.
Singapore’s Model AI Governance Framework has been adopted as a reference standard across Southeast Asia. Governance is happening here too, not just in Washington and Brussels.
None of this is glamorous. None of it goes viral. A researcher publishing a safety benchmark gets 200 LinkedIn impressions. A pundit predicting robot apocalypse gets 2 million views.
Bad news travels at the speed of fear. Good governance travels at the speed of bureaucracy.
But bureaucracy, for all its slowness, is how civilisations actually protect themselves. The seatbelt wasn’t invented by a viral moment. It was mandated — slowly, painfully, against industry resistance — and it has saved hundreds of thousands of lives since.
AI governance is at the seatbelt stage. Unglamorous, contested, incremental — and absolutely necessary.
Your job is not to panic and not to be naively optimistic. Your job is to stay informed enough to hold the right people accountable.
