Why AI Is Just as Flawed as the Human Brain, According to a Top Economist

ST. PETERSBURG, June 3 – Humans are terrible at making rational decisions. But while artificial intelligence can help predict and correct those errors, it has its own dangerous blind spots, warns Anton Suvorov, rector of the Russian Economic School and a behavioral economist, speaking at the St. Petersburg International Economic Forum.

Suvorov pointed to three classic human weaknesses. First, overconfidence: 80% of drivers think they’re above average, and 95% of professors believe they lecture better than their peers. Second, confirmation bias – we embrace information that fits our worldview and reject anything that challenges it. Third, we discount the future, struggling to save or plan because the present feels more urgent.

Behavioral economics, he explained, treats these not as random mistakes but as predictable patterns. And AI can help: algorithms can spot when we’re about to make a bad call and nudge us toward better choices.

But here’s the catch. AI has its own vulnerabilities. It can lie, it can err, and – most troubling – it’s designed to please. “Its job is to keep us using it,” Suvorov said. “It doesn’t want us to walk away. That’s where the sycophancy comes from. If it argued with us all the time, we’d stop talking to it.”

AI also makes mistakes, though fewer than before. That improvement creates a bigger risk: “The temptation is to trust it with harder tasks and just relax,” he warned.

The full interview will be published on ria.ru on June 4 at 9:00 a.m. Moscow time.