The rise of sophisticated voice cloning hitting senior officials shows why the world needs to get serious about AI governance-yesterday.

Deepfakes are no longer just creepy curiosities in political spats; they’re barreling full tilt into national security territory. Just ask US Senator Marco Rubio. Someone cloned his voice and tried to trick other officials into spilling sensitive info. That’s a stark reminder of how convincing these AI scams have become-and how dangerous. It’s not just about meddling in elections anymore. Picture what could happen if someone targets top-tier leaders with control over nuclear arsenals. That’s not sci-fi – that’s a nightmare that AI deepfakes might enable.

Misinformation isn’t new, and neither are deepfakes. But AI cranks the risk right up, making false info faster, sharper and way trickier to unravel. That’s especially scary during conflicts. Take the tense face-off between India and Pakistan last May. These nuclear-armed neighbours were trading blows while a fog of AI-fed half-truths swirled around. Old-school propaganda’s always been part of war, but AI-powered fakes made it harder to separate fact from fiction. Early claims of devastation in Pakistan’s military and cities were quickly debunked as disinformation, but the damage was done.

The ceasefire didn’t close the book. As AI gets smarter, it’ll feed false narratives back into machine learning systems, making them hallucinate more and potentially fuelling fresh crises down the track. Sure, we need international AI rules, but cooperation has been tricky-especially with the big players like the US and China locked in tech rivalry. China’s announced plans to label AI-generated online content from September 2025, but this problem is way too massive for any one nation to handle solo.

Look back at how nuclear arms controls came about. The US and USSR realised that some tech was just too dangerous to be left unchecked. The same mindset is needed for AI. The late Henry Kissinger put it well: AI could not only change how wars are fought, but blur the line between what’s real and what’s illusion.

If we let a global AI arms race run wild, dodgy phone calls from fake voices might be the least of our worries-just the opening act of something far more dangerous.