Philosopher Yuval Noah Harari warns of AI’s risks in finance

Avatar

The unbridled use of artificial intelligence in the financial system could lead to dire results, author, philosopher and history professor Yuval Noah Harari told an audience at the Bank for International Settlements (BIS) Innovation Summit. Strong institutions are needed to keep AI in check, he said.

The financial system creates trust, Harari argued. Money, bonds and other financial devices allow millions of strangers to work together toward common goals. But no effort has been made to make financial regulation understandable “to humans,” he said:

“Maybe 1% of the population understands how the financial system works. What happens when that number goes down to zero?”

Harari cited the financial crisis of 2007-2008 as an outcome of incomprehensible financial innovation. Regulators were unable to understand new financial products well enough to regulate them.

Related: Decentralized AI is key to more unbiased AI algorithms — Masa co-founder

AI is an alien form of intelligence that is only at the beginning of its evolution. It does not reason the way humans do and could easily create financial devices that are beyond human imagination, shifting power away from politicians and regulators to algorithms. Trust relationships would be built between AI systems rather than between people, and in a financial crisis, politicians and regulators would be forced to trust AI.

Yuval Noah Harari addressing an audience at the BIS Innovation Summit remotely. Source: YouTube

That disruption of trust relationships could lead to social instability. Trust in politicians and institutions is already declining, Harari noted, adding:

“We need to prevent AI from becoming completely unfathomable. […] We need the ability to understand it and regulate it.”

Regulation should not be left to geniuses or “charismatic leaders,” Harari continued. Only institutions can “keep humans in the loop.”

Source: Yuval Noah Harari

Even then, there are dangers. World-changing innovations are often successful only after a series of failed attempts to use them. Harari described imperialism, communism and Nazism as failed attempts to build industrialized societies — with tens of millions of victims. Meanwhile, AI is capable of coming up with ideas about how to use itself.

Harari said regulators need to concentrate on building more trust between people to ensure the future financial system is stable and human-friendly.

Magazine: What the hell is Web3 anyway?