�׻���������

Estimated reading time: 5 minutes

AI is a transformative technology. But does it serve all users equally? As the sector strengthens its influence in multiple domains, let’s explore how AI serves women and girls…

According to Forrester, global . But tech’s impact goes beyond economics. Today, it is embedded into every area of our lives. It shapes how we communicate, do business, relax, shop. 

Now, artificial intelligence is promising to take the impact of tech to new levels. AI systems are expected to replace or augment many human-centric tasks – from language translation to customer care to driving. McKinsey estimates Generative AI alone could add . 

This shift will change the culture too. So it’s critical to develop a ‘Responsible AI’ that serves all citizens. The foundations of AI tech are its training data sets – vast pools of information that AI systems use to create models of the world. But these models are not neutral. They reflect the world view, experience and world view of the employees, and people generating the data.

This raises questions around gender. Since most people working in AI are men, we might ask: do these models under represent women? Do their responses use stereotypes? Do they offer different advice based on the sex of the user?

As AI moves into the mainstream, societies are taking these Responsible AI questions seriously. Improving models is one way to address the issue. Another is to employ more women in the AI space since evidence suggests they are currently underrepresented.

Ultimately, achieving gender equity in AI will also make the tech better. AI systems that incorporate all human perspectives will deliver meaningful solutions to complex problems. Let’s dive into the facts and the arguments… 

Women in tech: the story so far

Many women have contributed to the history of tech – Yet still, in employment terms, women are in the minority. According to the World Bank, the number of women in STEM (science, tech, engineering and mathematics) jobs is 28.2 percent of the workforce.

This is a complex problem. While there may be some explicit (or more likely unconscious) bias against females, a more pressing problem is the supply line. Put simply, fewer women study tech. Indeed, research says women (in the US) account for . 

To address this issue, many corporations are taking action. In France, for instance, �׻��������� formed a partnership with the Elles Bougent Association to provide networking opportunities and mentoring for young women interested in STEM careers. Currently, more than 150 female �׻��������� employees are Elles Bougent mentors. 

Public bodies are playing a part too. The European Commission’s , for example, has helped more than 60,000 schoolgirls to pursue STEM degrees and take on leadership roles in the digital economy.
 

Women in AI: what’s at stake?

Is the female employment and training landscape any better in the AI space? It seems not. A study of 1.6 million AI professionals by Interface found that women . At senior levels, it’s just 14 percent.

However, the challenge of achieving Responsible AI goes beyond training and employment. An equally urgent task is addressing bias in the data sets and algorithms that train AI systems. 

These models reflect the perspectives and values of their creators. So when there is a lack of diversity in the workforce, it is bound to have an impact. There’s a risk that systems can mirror existing societal biases related to gender. This is an ethical concern – and women appear pessimistic about it. In a , 63 percent said they don’t believe that AI can be fully ethical in the next three years.

Exampled of bias have already been exposed. In 2018 . The tool looked for patterns in resumes and, since most came from men, it assumed downgraded applications from women. Meanwhile Nature revealed that the , which can lead to misdiagnosed conditions.

Generative AI: changing the workplace

With the launch of ChatGPT in 2023, a new form of AI – generative AI – hit the mainstream. GenAI is a leap forward from previous iterations of AI. Rather than identifying patterns or making predictions from existing data, it generates new content. Experts believe GenAI can transform industries and change the way we work.

There will certainly be job losses. But experts believe most jobs are ".” For this reason, it's important for women to learn GenAI skills. Are they? The data here is mixed. According to Deloitte, was just half that of men in 2023. However, it believes women are now adopting the tech 2.2x faster than men, and that both sexes will use the tech in equal numbers in some regions this year. 

Generative AI models: addressing the bias 

The challenge of bias in AI models is, if anything, even more pressing in generative AI than it is in older applications of the tech. GenAI systems use large language models (LLMs) to determine their outputs. The user writes a prompt, and the LLM gives an algorithmic answer using predictions based on its knowledge base. But, as stated, the algorithms and the data can be biased. 

Examples are already out there. When a asked image generating tools – DALL-E 2 and Stable Diffusion – to visualise roles like engineer and scientist, 75 to 100 percent of its generated images depicted men. Yet combating such bias is difficult. AI bias is hidden inside complex algorithms, making it hard to detect and address. 

Rising awareness helps. Tech companies are starting to confront the Responsible AI challenge in their products. For example, Google Translate now offers for some gender-neutral languages.

There is also action at the governmental level to boost inclusive policy-making. In Europe, the EU AI Act mandates that high-risk GenAI applications disclose their decision-making processes, include human oversight and use representative datasets.

Conclusion 

It’s clear that the AI space needs to do more to address the gender disparity in its workforce, and the bias in its training sets. AI tools, arguably more than any other tech, must reflect all perspectives in society in order to generate accurate outcomes and maintain trust.  Male AI leaders and employees can play a role by addressing the unconscious bias that might be degrading the quality of AI models. In so doing they will help to create a world of Responsible AI that will benefit everyone, not just women.