Collective Intelligence — or Algorithmic Herding?

This reflection extends Behaviour & Responsibility — the systemic risk that emerges when a society thinks with the same tools.

Artificial intelligence can help groups learn from more information. It can combine expert knowledge, identify patterns across organisations and share successful practices rapidly. This may create a form of collective intelligence.

But when many people rely on the same models and recommendations, they may also begin making similar decisions. Investors may follow similar signals. Companies may adopt similar strategies. Creators may produce similar styles. Job applicants may use similar language.

What appears to be independent judgement may originate from the same underlying system. This can create algorithmic herding.

Uniform behaviour may initially look rational because each person follows a sophisticated recommendation. But when everyone acts similarly, systems can become fragile. A market may move sharply because many automated strategies respond to the same signal. Businesses may overlook an opportunity because their models share similar assumptions. Public discussion may narrow because AI-generated summaries repeatedly prioritise the same interpretations.

Diversity of judgement is therefore not inefficient noise. It can protect society from common error.

Organisations should avoid depending on a single model, dataset or framework for every important decision. Human teams should examine alternative interpretations and ask which assumptions may be shared across systems.

AI can help coordinate knowledge, but coordination should not eliminate independent thought. The strongest collective intelligence does not require everyone to reach the same conclusion. It allows different perspectives to challenge one another before a shared decision is made.

The intelligent age should be careful not to confuse agreement produced by common tools with agreement produced through genuine understanding.