Twelve years later and "man" is still more similar to "doctor" than "woman"
Despite advancements over the last decade, language models like Gemini still exhibit ingrained biases. This persistent issue stems from AI’s fundamental reliance on vast, inherently biased datasets, hindering the creation of truly reliable and sensible systems.