The 21st century is currently marked by mankind’s very own digital revolution. Despite most people considering it as beneficial, whether it spells betterment or not remains to be seen. It has however proven to be a threat to that which encompasses all sorts of aspects of human societies ; to people’s cultures, instead of their preservation. As AI tools users, you should be capable of distinguishing what is contextual and what is statistical, what should be researched and what should be transmitted. The boundaries between AI and culture are clear and solid.
Incompatibility in essence
Origin story
The term « artificial intelligence » was offcially first heard in Dartmouth College. McCarthy and his colleagues adapted it in the 50s for their event «The Dartmouth Summer research project on Artificial intelligence ». This project represented their greed to emulate some aspects of human thoughts through machines. However what people are becoming heavily dependent on today, is but an amalgamation of calculations and data. And this variety of programs have not been swallowed and mutated to adapt to their culture like most advancements.
Cores of AI and Culture
What belongs to the concept of culture, are answers to :
- Who we are
- What is appropriate
- Where we are
- What makes us different
- How we communicate.
AI answers: «What is the most likely response?» ; it is only about function. Those goals don’t fully overlap, thus AI cannot be a culture on its own.
The reason why culture is vague, is because it heavily depends on:
- Shared history
- Geography
- Social norms
- Implicit meanings
- What is not necessarily said but acted upon
AI, on the other hand, works by detecting patterns in data, so it may know that something is common, but not why it matters, like how it cannot comprehend all sorts of idioms that differ even in regions within the same country. Hence at this moment, we do not recommend relying on AI for tasks that require cultural understanding or a cultural context.
AI do not live culture
Humans experience culture through: body language, space, emotion, daily routines. Meanwhile AI has no body, no social risk, no lived experience. It can describe a ritual, but it can’t feel embarrassment, pride, or discomfort — all crucial cultural signals.
Each cultural perspective is limited and particular with their ideologies and prejudices. Whereas most AI systems are trained primarily on western english-dominant, online-visible cultures. This means: Minority cultures get simplified, non-dominant norms get misinterpreted, «Neutral» AI often reflects hidden cultural bias, result of provided data.
As such AI is not a culture on its own but simply a result of digital revolution much how steam machine were the results of industrial revolution, it is a tool coming with benefits but equal or graver harms to those who use it and those who don’t.
The TCLoc master can give proper guidance on the ethical use of AI, be it in localization or technical communication with proper respect to culture and its languages.
