Ex-Google CEO’s 2026 warning: AI will be smarter & uncontrollable

Former Google CEO Dr. Eric Schmidt recently shared some eye-opening predictions about artificial intelligence that should make us all sit up and pay attention. Speaking at the Special Competitive Studies Project, Schmidt suggested that we’re drastically underestimating how quickly AI is evolving toward superhuman capabilities.

AI is underhyped, not overhyped

While many tech observers complain about AI hype, Schmidt argues the opposite: we’re not taking the implications seriously enough. Why? Because we’re potentially just six years away from artificial superintelligence (ASI) – systems smarter than the collective intelligence of all humans combined.

“This path is not understood in our society. There’s no language for what happens with the arrival of this,” Schmidt warned. “This is happening faster than our human society, our democracy, our laws will address.”

The coming AI revolution in programming

One of Schmidt’s most striking predictions concerns the future of programming. Within just one year, he believes AI will be writing nearly all code. This isn’t just Schmidt’s view – Dario Amodei of Anthropic made similar predictions, suggesting we’re only 3-6 months away from AI writing 90% of code, and within 12 months, “AI may be writing essentially all of the code.”

But what happens next is even more fascinating. Schmidt points to something called “recursive self-improvement,” where AI systems begin improving themselves without human intervention. Major AI labs already report that 10-20% of their code is being generated by their own AI systems.

The San Francisco consensus: AGI by 2026-2028

What Schmidt calls the “San Francisco consensus” suggests we’ll achieve artificial general intelligence (AGI) within 3-5 years – systems as smart as the brightest humans across multiple domains simultaneously.

“What happens when every single one of us has the equivalent of the smartest human on every problem in our pocket?” Schmidt asked. ”