Skip to content
THOUGHT
Apr 24, 2026
Pontus Wadström, PhD in Strategy and Strategy Advisor
Pontus Wadström
PhD in Strategy & Strategy Advisor

Why AI bias Isn't necessarily a bad thing in the context of strategy

When we discuss AI, the word "bias" usually sounds bad. It suggests prejudice or unfairness. However, bias exists in both people and AI models, but it comes from different places. For us humans, it's down to the shortcuts our brains take and our life experiences. For AI, it's simply what it learned from its huge training datasets. However, what if we looked at 'bias' in AI differently? Instead of just seeing it as a fault we must get rid of, what if we used it as a tool for deliberate perspective-taking?

From Flaw to Feature: The Power of Different Viewpoints

We are presuming an AI that doesn't just give one answer, but a range of informed opinions straight away. We are not talking about creating harmful prejudice; we are talking about designing AI to clearly express specific, thoughtful positions—just like when diverse human experts give varied analyses.

This approach accepts that true understanding often comes from putting different views together. A "biased" model, in this positive sense, isn't faulty. It's purposefully designed to represent one specific viewpoint or area of knowledge or a point of view.

Multi-Agent Systems: The Architecture for Diverse Views

AI that mixes different types of information, like text, pictures, and sound (multimodal AI), is much better at understanding the full picture. When we link this to multi-agent LLM systems (Large Language Model systems where several specialised AIs work together), we create a powerful system for generating diverse perspectives.

Say a tricky public policy issue is being discussed. Instead of one general-purpose AI, a multi-agent system could deploy several specialised agents:

  • An "Economic Impact" Agent: An AI that focuses only on money matters and financial data.
  • A "Social Equity" Agent: An AI that focuses on fairness, ethical rules, and how society is affected.
  • An "Environmental Sustainability" Agent: An AI that focuses on climate data and the impact on nature.

Each agent, using its focused "bias" or specialisation, looks at the same information but through its unique filter, generating distinct and well-reasoned analyses. These separate outputs then come together. A human analyst can then compare and mix these views to get a far more comprehensive and rounded understanding than any single, general AI could provide.

Cultivating a Vaster Knowledge Vector

Using these distinct AI perspectives on purpose gives us profound advantages:

  • Better Human Decisions: People get a clear, structured breakdown of the problem from several angles. This makes us think harder and leads to sounder decisions.
  • Fewer AI Blind Spots: By cross-checking what the different "biased" models say, the system can spot gaps, things that don't quite add up, or areas where just one perspective might be misleading.
  • Quicker Problem Solving: For complex research, design, or strategic planning, specialised AI agents can quickly explore lots of potential outcomes from their distinct points of view.

From this modern perspective, bias is not a problem to be wiped out, but a feature to be carefully built and used. By intentionally crafting models with specific viewpoints and world views based on design choices and the views of its creator and that data it has been trained on we could empower both machines and humans to construct a truly vast and detailed body of knowledge. The future of AI isn't just about general intelligence; it's about cleverly curated, diverse intelligences working in concert.

The Essential Partnership: Hybrid Intelligence

This brings us to the final aim: not replacing people, but achieving hybrid intelligence, the deep partnership between humans and AI. When these specialised AI agents give their distinct, multi-faceted views, they hand over a much richer and more structured set of information to the human strategist or analyst. The human job is then to apply their unique judgement, moral framework, and years of lived experience to synthesise these viewpoints and make the final, informed decision. The truly vast knowledge vector, and the best strategic outcome, is achieved only when clever, specialised AI works hand-in-hand with irreplaceable human expertise.