Artificial intelligence agents are beginning to interact in ways that create risks beyond individual misalignment with corporate and social expectations. As happened with global finance after the 2007 crisis, AI governance needs to begin focusing on how good agents can still produce bad systems.

The discussion about concentration in artificial intelligence markets focuses on the least concentrated layer, the models. The chokepoint that actually threatens AI is the production of refined minerals that go into chips, data centers, and electricity production, writes Piyush Akimitsu.   

Walid Chaiehloudj argues that when a scientifically substantiated large but uncertain risk, like environmental damage, conflicts with standard competition analysis, competition authorities need to defer to a precautionary citizen-consumer standard that takes into consideration the risks of that harm.

In recent research, Jorge Alé-Chilet, Cuicui Chen, Jing Li, and Mathias Reynaert find that when faced with environmental regulations, collusion among German car manufacturers reduced their expected non-compliance fines and significantly increased consumer and producer surplus. At the same time, social welfare decreased by billions of euros because of increased pollution.

Wealth taxes can make capital markets more efficient when they are optimally combined with lower capital gains taxes, argue Sergio Ocampo, Guttorm Schjelderup, and Floris Zoutman in new research.

Anticompetitive behavior, such as consolidation, in the health-insurance industry has allowed a small group of insurers to dominate the market. One of the main drivers of this behavior is “adverse selection,” where individuals know more about how sick they are than insurance companies do. To mitigate the risk of paying out more claims than they can cover, insurers often raise their premiums. Kellogg’s Amanda Starc and her colleague argue that, to create a truly healthy health-insurance marketplace, regulators and policymakers need a framework for rethinking the complex interactions between adverse selection and regulatory guardrails.