Working Together AP-1.2

Protect Our Cultures

AI should celebrate our differences, not flatten them.

Your grandmother had a word for the way morning light hits the kitchen table. A word that exists only in your language, carrying centuries of meaning. Now imagine an AI trained mostly on English deciding that word is unnecessary — translating it away, smoothing it out, replacing it with something generic. Piece by piece, the world starts to sound the same.

What This Means

AI systems are trained on massive amounts of data, and most of that data comes from a handful of dominant languages and cultures. This means AI can accidentally erase smaller languages, flatten local traditions, and make everything feel like it was made in Silicon Valley. This policy demands that AI respect and actively preserve the beautiful diversity of human culture — every language, every tradition, every unique way of seeing the world.

A Real-World Scenario

A school in Wales started using an AI tutoring system to help students with their homework. But the system only worked well in English. Welsh-speaking students found their essays "corrected" to remove Welsh idioms and expressions. The AI treated their cultural identity as a grammar mistake. After parents raised concerns, the school demanded an AI system that actually supported and celebrated bilingual education.

Why It Matters to You

Because your culture is not a bug to be optimized away. The food you grew up with, the music that moves you, the way your community tells stories — these things make life rich and meaningful. If AI homogenizes everything, we all lose something irreplaceable.

For the technically inclined

AP-1.2: Cultural Diversity

AI systems should preserve and promote cultural diversity rather than homogenize cultural expression, language, or creative output.

What You Can Do

When you use AI tools, check if they support your language well — not just technically, but culturally. Report cases where AI flattens or disrespects cultural content. Support organizations working to include minority languages in AI training data.

Related Policies

Stay Updated

Get notified about new policies and AIPolicy updates.

No spam. Release updates only.