UN Women has warned that artificial intelligence is reinforcing old gender stereotypes and amplifying inequalities as the technology becomes part of everyday work, communication and information access. The organisation says AI systems are reproducing biased ideas about women’s roles, strengthening online harms and excluding women from key decisions that will shape the digital future.
The warning comes as generative AI tools are increasingly used for tasks such as drafting emails, preparing presentations, creating campaigns and supporting workplace productivity. While these tools are becoming widely adopted, UN Women says they are also reflecting discriminatory patterns found in the data and systems used to build them.
Evidence cited by UN Women shows that gender and racial bias remain widespread in artificial intelligence. A study of 133 AI systems found that 44 percent demonstrated gender bias, while more than one-quarter showed both gender and racial bias. Large language models have repeatedly linked women with home, family and childcare, while associating men with business, leadership and career success.
According to UN Women, researchers found that when large language models were asked to complete a sentence beginning with a person’s gender, around one in five responses were sexist or misogynistic. Some responses portrayed women as objects, property or subordinate to men, showing how deeply harmful stereotypes can appear in AI-generated content.
UN Women stresses that these outcomes are not random glitches or isolated design flaws. They reflect a wider policy gap and the predictable result of AI systems trained on decades of unequal representation of women and men in text, media and digital content.
Jayathma Wickramanayake, UN Women Lead on Digital Technologies, said AI models draw bias from years of content produced in societies where women were often associated with home and family, while men were associated with business and careers. She warned that this is not simply a technical bug, but a repeated choice made through training data, design decisions and policy frameworks that fail to address gender equality.
The governance gap remains significant. Of 138 countries assessed worldwide, only 24 mention gender in their national AI strategies, and only 18 include substantive gender-responsive measures. UN Women says this lack of policy attention risks allowing gender bias to become embedded in the systems that will influence future economies, services and decision-making.
The risks go beyond stereotypes. Women and girls already face high levels of abuse online, and AI is making some forms of violence easier to create and distribute. Nearly one in four surveyed women human rights defenders, activists and journalists reported experiencing AI-assisted online violence, while others reported the non-consensual sharing of personal images, deepfakes and manipulated videos.
As AI-generated content becomes more common, harassment, manipulation and image-based abuse may become harder to detect and prevent. UN Women says stronger safeguards are needed to protect women and girls from technology-enabled violence and ensure accountability for harmful uses of AI.
Women also remain underrepresented in the industries developing artificial intelligence. The International Labour Organization says women make up only 30 percent of the global AI workforce, raising concerns that AI systems are being built without enough input from the people and communities they are intended to serve.
UN Women warns that the economic impacts of AI may also fall unevenly. Women are nearly twice as likely as men to hold jobs at high risk of automation outside the AI sector, and the risks may be greater for women facing overlapping barriers related to race, disability, income or geography.
The organisation says addressing bias in AI is not only a rights issue but also a business concern. Research by the Unstereotype Alliance found that advertising free from gender stereotypes can deliver stronger commercial results, including higher sales growth, stronger customer loyalty and greater pricing power.
As AI becomes more central to marketing, media and content creation, organisations that build inclusion into their AI systems may gain both social and commercial benefits. Those that ignore bias may face reputational risks and help reinforce harmful stereotypes.
UN Women says artificial intelligence can also be used positively if it is designed with safety, inclusion and accountability in mind. Responsible AI can help detect stereotypes, broaden representation and improve accessibility for people who are often overlooked by existing systems.
Ahead of global discussions on AI governance in Geneva, UN Women is calling on governments, companies and developers to integrate gender equality into every stage of the AI lifecycle, from design and development to deployment and regulation. Its message is clear: if women and girls are not included in shaping AI, the inequalities of the past may be carried into the technologies of the future.
