Report Highlights Western Bias in AI
A report by Chromatics AI, as covered by ANI, Livemint, and The Tribune, warns that Western cultural bias built into artificial intelligence systems could influence how billions of users understand and interact with the world. As AI tools increasingly become part of areas such as education, professional counselling, and legal guidance, the report says these embedded values risk shaping how under-represented users perceive the world around them in ways that may be directly at odds with their own.
The report, dated August 24, states that many AI systems are trained on data that does not adequately represent the languages, worldviews, and ways of communication of large parts of the global population. As a result, AI responses can appear neutral while still reflecting cultural values associated with the West. 'Research suggests that leading AI systems exhibit a default prioritisation of individualism and Anglo-Saxon norms,' the report says.
The Limits of Localization
The report argues that simply translating AI systems or adapting them for different regions may not be enough to address the problem. Localization, it explains, generally focuses on surface-level changes such as translating an interface, changing currency symbols, or applying regional style guides, while the underlying AI model continues with the cultural assumptions from its original training data.
This creates, per the report, a gap between language localization and genuine cultural understanding. Users from communities historically under-represented in technology, it says, can notice when an AI system misunderstands their context, communication style, or values.
Examples of Culturally Grounded AI
To illustrate efforts to build AI rooted in local languages and contexts, the report points to Egypt's Horus, Nigeria's N-ATLaS, Indonesia's Sahabat-AI, Latin America's Latam-GPT, Thailand's OpenThaiGPT, and AI Sweden's Swedish-language model. It also cites Amazon's work on Alexa in Mexico, where, as reported by Livemint and The Tribune, the company used reinforcement learning and custom training data to adjust for language and country while ensuring improvements did not harm other languages.
Founder's Perspective
Larry Adams, Founder of Chromatics AI, stressed that bespoke culture AI needs resources not yet prioritized at scale. As quoted in Livemint, Adams said, "Training data must be collected in genuine partnership with the communities it is meant to serve. Evaluation benchmarks must be culturally grounded, not just linguistically accurate." He added, "Communities themselves should help define what correct looks like."
The report concludes that the next phase of AI development may depend not only on larger models, but also on how effectively systems understand the cultural contexts of the people using them.