The IMF's second-in-command recently warned that the true disruptive effects of artificial intelligence on the economy and financial markets may not become apparent until there is a recession, which could turn into a full-blown crisis. Unless the risks of AI are addressed.
During a speech at an AI summit in Switzerland on May 30, IMF First Deputy Managing Director Geetha Gopinath said the debate about the risks of AI has mostly focused on privacy, security and disinformation. But little has been said about the threat of AI fueling the next recession.
In a world of widespread AI adoption, the technology could turn an otherwise modest recession into a much deeper economic crisis by disrupting labor markets, financial markets and supply chains, he said.
AI Threats to Labor Markets
In normal economic times, companies have historically tended to invest in automation but still hold on to workers because they benefit from doing so. But when companies cut costs in a recession, workers are laid off and replaced by automation, he explained.
Gopinath points to IMF research showing that in advanced economies, 30% of jobs are at high risk of AI substitution, compared to 20% in emerging markets, and low-income countries. I 18%.
“So we have a very wide range of potential job losses that we could have,” he warned. “And again, the risks of long-term unemployment are quite acute.”
Risks of AI in Financial Markets
The financial industry has long embraced automation and early forms of AI, such as algorithmic trading, and the sector is rapidly adopting new AI technologies today.
Gopinath noted that some AI trading is being replaced by more complex models that can learn on their own, and forecasts suggest that robo-advisors will control more than $2 trillion in assets by 2028, up from $2 trillion in 2023. Less than $1.5 trillion.
He added that while AI can improve market efficiency and inclusion, the risks of AI are also more likely to be reflected in the downturn. This is because new AI models will perform poorly on new events that are different from what they were trained on.
“And one thing we know is that no two recessions are the same,” Gopinath said.
In such a scenario, AI could encourage a rapid, simultaneous move to safer assets, which in turn lowers riskier asset prices.
The AI models will then detect a drop in price, see that as a validation of their initial moves, then double down with more asset sales. And given the black-box nature of AI, such behavior can be difficult to control.
“You can have fire selling and losing behavior, which leads to even bigger falls in asset prices,” Gopinath said.
AI Risks in Supply Chains
As businesses adopt AI, they let it play a bigger role in deciding how much inventory to keep and how much to produce.
In normal economic times, this can increase efficiency and productivity. But he said AI models that were trained on “stale data” could make big mistakes and lead to supply chain disruptions.
Ways to mitigate AI risks
After presenting the dire scenarios, Gopinath also provided recommendations for mitigating the risks of AI without undermining the positive side of AI.
One way is to ensure that tax policies don't inefficiently favor automation over workers, though she was careful to note that she's not proposing a special tax on AI.
Another approach is to strengthen the social safety net with more generous unemployment benefits, along with worker education and new skills.
He added that AI could also be part of the solution, such as upskilling, better targeting of aid, and flagging early warnings in financial markets.
“I believe there is a real need to make a parallel effort to ensure that we AI-proof the global economy as well,” Gopinath said.
His warning comes a year after he said we don't have much time to determine how to protect people from AI.
“We need governments, we need institutions and we need policymakers to move quickly on all fronts, in terms of regulation, but also in terms of preparing for potentially substantial disruptions in labor markets.” . Financial Times.