What Is AI Herding?

AI herding happens when organisations use similar AI systems to make decisions and begin reaching the same conclusions. This can be useful for consistency, but it also raises a concern: If everyone relies on the same technology and information, we could start losing the different opinions and approaches that normally influence business decisions.

Take insurance, for example. Different insurers might look at the same business and reach different conclusions about whether to offer cover or how much to charge. Their decisions depend on their experience, the information available and how they assess the risks involved. But if several insurers start using similar AI systems, those differences could gradually disappear. The concern is what happens when this starts affecting an entire industry;

• Banks might become less willing to lend to certain businesses
• while insurers could begin turning away the same types of customers.

Each organisation might believe it’s making a sensible decision, but together those decisions could create a much bigger problem. AI herding isn’t necessarily about AI getting something wrong, but about what happens when too many organisations start reaching the same conclusions.