WHEN BIAS PASSES FOR NORMAL
How the perspectives most often represented in the data can become the norm — and how AI can fill in missing information, smooth over differences or confirm our hypotheses.
LA DUPLA’S POINT OF VIEW
BIAS,
WELL WRITTEN.
When AI fills in the gaps in our knowledge.
A convincing answer can still hide an assumption. A tool to help you examine bias and cultural representation in your everyday work with AI.
DOWNLOAD THE TOOL ↙
WHAT YOU WILL FIND
BEFORE YOU ACCEPT
AN ANSWER.
THE 3Cs IN ACTION
Curate your sources, create your instructions and check the interpretations. Three stages for examining what we know, what we assume and whose voices may be missing.
Adapted from the Unstereotype Alliance’s 3Cs Framework (UN Women).
INSTRUCTIONS FOR YOUR NEXT PROMPT
Seven instructions to preserve nuance, distinguish evidence from hypotheses and identify blind spots. An extra layer of critical judgement for your everyday work with AI.
HOW TO USE IT
YOUR USUAL TASK.
DIFFERENT QUESTIONS.
Make your usual request to the AI, attach the necessary materials and add the instructions from the document. They work alongside the task you already wanted to complete.
The tool does not replace a researcher’s judgement: it cannot correct a biased sample or bring back voices we never heard.
DOWNLOAD
THE TOOL.
Enter your email address to download the PDF, including the framework, the 3Cs and the complete set of instructions to add to your AI prompts.
