From the course: The AI Equity Imperative: Building a More Inclusive Future with AI
Unlock this course with a free trial
Join today to access over 24,700 courses taught by industry experts.
The promise and challenge of AI
From the course: The AI Equity Imperative: Building a More Inclusive Future with AI
The promise and challenge of AI
- The global reach of artificial intelligence has exposed critical disparities between who benefits and who doesn't benefit from these powerful technologies. English dominates the internet, representing at least 50% of all web content. However, according to Ethnologue, only 18% of the world's population speaks English. This disparity creates a fundamental imbalance in AI training data, leading AI models to disproportionately represent English. This means that mainstream language models are not serving languages with hundreds of millions of speakers, like Hindi, Urdu, Bengali, Indonesian, and Arabic. The media number of languages served by prominent LLMs like Gemini, ChatGPT, Claude, Llama, AYA, Falcon, and DeepSeek is just 12, with the average being 33. These multilingual LLMs also vary in terms of performance, often favoring English and other Western languages like French, Spanish, Portuguese, and Italian, again, once more ostracizing a significant portion of the world's population…