A 2024 UNESCO study of 133 AI systems found that 44% showed clear gender bias, slotting women into domestic roles four times more often than men; feeding a woman’s name produced “family” and “children,” while a man’s name yielded “career” and “salary.” A separate 2024 study found AI models score a CV more favourably when a man’s name sits at the top, with the same qualifications and experience. In 2019, Apple Card’s algorithm handed out higher credit limits to men than women, even when women had identical or stronger credit histories. In 2018, MIT and Stanford found facial recognition software’s worst failures fell on women of colour and its best results on white men. In 2015, Amazon’s AI recruitment tool, trained overwhelmingly on men’s CVs, learned to prefer men. Research from the LSE in 2025 found AI tools used by the majority of English councils systematically played down women’s physical and mental health problems compared to men’s.