23/06/2026
AI systems do not see the world.
They operate within a limited field of perception defined by the data they receive and the objectives they are asked to optimize.
This is the second of five observations explored in Human vs. AI, a report by With Company based on a parallel experiment comparing human-led and AI-supported approaches to the same creative challenge.
What lies outside that field simply does not exist for the system.
And yet, this is often where the most important externalities live: social impact, long-term consequences and unintended effects.
AI can optimize extremely well within a defined frame.
But defining that frame remains a human responsibility.
For organizations, the challenge is not only improving what AI can see. It is becoming more intentional about what matters, what should be optimized, and what risks being excluded from the system’s field of attention.
👉 Download the full report in the link in bio.