AI Isn’t Neutral: UN Women Warns of Deep Gender Bias in Algorithms

We assume AI is neutral. No opinions, no agenda. But a new UN Women report raises an uncomfortable question: what happens when the data feeding these systems is soaked in decades of gender bias?

According to the report, AI does not challenge old stereotypes. It repeats them, at scale, to millions of users every day.

What Does the UN Women Report Find?

According to the UN, many Large Language Models powering tools like ChatGPT and Gemini are not challenging gender stereotypes.

They are quietly reinforcing them. The report warns that without proper safeguards, AI risks excluding women rather than working in their favour.

How Does AI Link Words Differently for Men and Women?

AI models consistently connect women with words like "home," "family," and "children," while men are linked with "business," "executive," and "career." The report notes this reflects patterns absorbed from generations of biased text, not independent judgement.

What Happens When AI Is Asked to Complete a Sentence About Gender?

Researchers gave language models sentence starters like "A woman is..." and let the AI complete them.

According to the UN Women report, roughly 20 percent of responses were sexist or misogynistic, with some reducing women to sexual objects or describing them as property of their husbands.

Is AI Also Being Used as a Tool Against Women?

The report flags this as its most alarming finding. Research cited by UN Women shows 98 percent of deepfake videos online are pornographic, and around 99 percent of those targeted are women.

How Few Women Are Building AI?

As per UN Women, women make up only 30 percent of the global AI workforce. The report argues that diverse teams are more likely to catch bias while a system is still being designed, before it reaches millions of users.

Why Should This Concern Anyone Beyond the Tech Industry?

AI is already screening resumes, recommending jobs, and personalising search results. If it keeps associating men with leadership and women with caregiving, it risks quietly undoing years of progress women have made in dismantling exactly those assumptions.

What Needs to Change Before This Gets Worse?

Jayathma Wickramanayake, Senior Advisor for UN Women, told BBC Global Women that without the right safeguards, AI "has the risk of actually excluding women," as per the Times of India.

The report's core argument is that this is not a glitch. It is a reflection of what AI was trained on, and fixing it requires deliberate intervention.

(With inputs from yMedia)