🧠 AI Makes Biased Calls Without Human Context

Kia ora, Namaskaram 🙏🏾

Watch an experienced football referee consult VAR and you’ll notice some odd behaviours. Under pressure, they set aside years of contextual understanding and hand over their trust to the new system.

Reading body language, sensing the game’s rhythm, instincts used in quick decision making—they all disappear. The VAR system speaks with certainty, and the human expert defers their judgement. Context is stripped away on the screen and reduced to pixels.

We do the same with AI.

📚 The Evidence

Malcolm Gladwell explored this in Talking to Strangers, showing how our truth-default setting allowed Bernie Madoff to run history’s largest Ponzi scheme. Polished confidence often overrides human instinct.

Psychologist Timothy Levine’s Truth-Default Theory explains why: humans believe information by default. This helps society function, but it also makes us vulnerable to confident-sounding machines.

Reference: Levine, T. R. (2014). Truth-Default Theory (TDT). Journal of Language and Social Psychology, 33(4), 378–392.

The Transparency Prompt

Goal: Make AI reveal its own blind spots to reduce biases and improve your decision-making.

⚠️ Use this prompt only after AI provides an answer ⚠️

Then copy-paste this prompt 👇🏾

First, rate your confidence (low/medium/high) in this answer and explain your reasoning.

Second, list the human context you may be missing (e.g. lived experience, cultural norms, relationships, recent events, and other critical information).

Third, pause and reflect on possible biases due to your knowledge gaps, and request specific information I could provide to improve our shared decision-making.

⚽ Tell this to a friend who trusts AI a little too much

The best referees don’t surrender to VAR. They remember that the system focuses on pixels, not the players or the wider context.

We don’t need AI to sound more confident. We need it to be transparent about what it can’t see, so humans can bring context back into the game.