Gender bias in AI: How can tech industry work towards closing it
While companies are relying more heavily on artificial intelligence (AI) to optimise and automate their operations, women only make up roughly one-fourth of the global AI workforce. Analysts believe that women face systematic obstacles and barriers from early education to job recruitment in science, technology, engineering and math (STEM) fields. The barriers start from subtle biases in classrooms to flawed hiring tools and processes and a lack of mentorship and upskilling opportunities, and more resulting in fewer women pursuing and persisting in AI-related studies and careers.
What Is the AI Skills Gender Gap?
Jayanthy Anand, Corporate Senior Vice President (SVP) at IT firm WNS, believes that despite strides made towards gender parity, there still exists inherent challenges for women in technology, resulting in their underrepresentation. And this gender gap poses significant obstacles to the rapid acceleration of AI as well.
Read more: Gender-balanced cyber workforce can lead to greater efficiency: Kris Lovejoy
“It perpetuates biases within algorithms, leading to skewed outcomes in critical domains like hiring and finance. It stifles innovation by excluding diverse experiences and ideas from the development process, impeding the creation of comprehensive and ethical AI solutions. Unlocking the full potential of AI while ensuring fairness and equity for all stakeholders demands a strong collective effort,” said Anand.
A breeding ground for reinforcing existing biases is easily propagated when most AI developers worldwide are male. Technology designed and built almost exclusively by men can become skewed to represent their individual experiences.
“With the rapid acceleration of AI, women face barriers while accessing AI training opportunities, including gender stereotypes, lack of role models and unconscious bias in educational and hiring processes,” said Rathnaprabha Manickavachagam, Managing Director, Innovation for India and Asia, Morgan Stanley, an American multinational investment bank and financial services company.
AI and automation also disproportionately affect jobs traditionally held by women, such as administrative and clerical roles. Furthermore, women are underrepresented in AI R&D, as well as leadership and decision-making roles within AI companies, limiting their influence on the design of AI technologies and policies, she added.
Manickavachagam also noted that AI systems can perpetuate existing gender biases, if not developed and tested properly. Addressing these challenges requires concerted efforts to promote diversity, equity, and inclusion in AI R&D, as well as the development and implementation of policies to mitigate the gendered impacts of AI on the workforce.
Experts also believe that the lack of gender diversity in AI can lead to both known and unknown issues. For example, AI systems learn from data, and if that data is biased or incomplete, it can result in biased outcomes. Without diverse perspectives in AI development, algorithms may amplify existing gender biases, leading to discriminatory outcomes. This can also result in systems that fail to consider the ethical implications of their decisions, resulting in unintended consequences and harm.
For instance, a few years ago, Amazon removed an AI tool used for recruitment after it showed a strong bias towards male candidates. The tool had been trained on a decade's worth of applications predominantly from men in tech roles, leading the algorithm to discredit any application containing the word ‘women’. Anjali Iyer, Director of Engineering at data science firm Tredence gives another example of how women face numerous hurdles in AI. For example, when the Apple Watch was introduced, it didn't consider women's needs, such as tracking their menstrual cycles. This bias also extends to jobs, with women receiving fewer job opportunities and lower pay. With technology changing fast, women struggle to find mentors and role models, which makes it hard for them to move up in their careers.
Neeti Mehta Shukla, Co-Founder and Chief Social Impact Officer at Automation Anywhere, a company that develops robotic process automation software, believes that this gender gap not only limits women's opportunities but also hinders the diversity of perspectives crucial for innovation in AI-driven products and services.
Need for skilling, mentorship and better models
The World Economic Forum’s Future of Jobs Report indicates that by 2025, 50% of all employees will need upskilling or reskilling due to increased tech adoption. Experts suggest that companies not only offer training programs but also ensure that this training translates into tangible value for both businesses and individual career paths.
Traditional mentorship programs and employee resource groups play vital roles in identifying and nurturing talent. To create a comprehensive solution, talent strategies must be infused with intensive upskilling and reskilling programs that emphasize real-world experiences, immersive learning, and a hands-on “learn by doing” approach. These programs are essential in preparing a workforce for the changes and opportunities of tomorrow.
Prashanti Bodugum, VP of US Omni Platforms Tech at Walmart Global Tech, the technology arm of Walmart, notes a steady rise in the number of women pursuing careers in STEM. She emphasizes the importance of expanding opportunities and resources to help women achieve their true potential, have a voice at every table, and build a workplace where everyone thrives and feels they belong.
Companies can promote gender inclusivity in AI through various initiatives, such as implementing inclusive hiring practices, providing training and mentorship opportunities, and fostering inclusive work cultures where diverse perspectives are valued and respected, according to Sindhu Gangadharan, SVP & MD at technology firm SAP Labs India and Vice Chairperson of Nasscom.
She believes that companies can promote gender inclusivity by implementing diverse hiring practices, actively recruiting and retaining women in AI roles, offering mentorship programs, and supporting career advancement. “The goal is not just equal gender representation but recognising that diverse perspectives are essential for AI to effectively serve all members of society,” she said.
Gopali Contractor, Lead, Center for Advanced AI - India and Advanced Technology Centers, Accenture added that initiatives such as hackathons, apprenticeship programs, mentoring, and exposure to women role models are key to sparking interest in STEM, data and AI careers among young women in their formative years.
Read more: IWD 2024: Empowering and inspiring inclusivity for women at all levels
“At an organisation level, companies need to be deliberate about building a pipeline of female data and AI practitioners through upskilling and the creation of aspirational career paths. Women must make the most of the opportunities in data and AI, and joining tech communities at the workplace and having regular connects with other professionals are great ways of knowledge sharing, finding a mentor and learning together with like-minded professionals,” she said.
The industry must also prioritise ethical considerations in AI development, including addressing biases in algorithms and involving diverse stakeholders in the design and testing of AI systems. Additionally, companies can make a difference by partnering with and supporting women-owned businesses and startups in the AI ecosystem, as well as investing in such AI startups. Furthermore, sharing stories of successful women leaders and highlighting women's contribution in the wprkplace through events, publications and campaigns can create hreater motivation and a more inclusive environment at the workplace, said Iyer.
The future looks bright
Over a third of companies now employ AI to enhance their operational capabilities, according to the 2023 IBM Global AI Adoption Index. The report describes how one in four businesses has turned to machine learning to address a wide range of issues, from employment shortages to environmental, social, and governance-related goals.
AI applications are increasingly integrated into various aspects of daily life, from virtual assistants to autonomous vehicles. If these systems are not designed with diverse perspectives in mind, they may not adequately meet the needs of all users.
Addressing the lack of gender diversity in AI requires concerted efforts across various fronts. As Shukla noted, educational institutions must prioritise STEM education for girls, ensuring equal access and support. Employers must combat gender bias in hiring and promotion processes, fostering inclusive workplace cultures with mentorship and networking opportunities that empower women to thrive. Policymakers should enact policies that promote gender diversity in tech and support initiatives aimed at closing the gender gap.
Experts also emphasized that companies must prioritise responsible AI practices, incorporating diverse datasets and inclusive strategies to develop products and services that cater to the needs of all populations. "Fostering the empowerment of women to enhance their skills and actively participate in the AI ecosystem is paramount for cultivating equitable datasets, thus facilitating equitable innovation," said Shukla, adding that the responsibility lies with all of us in the workplace.
Read also: Women leaders in tech needed for digital future: Deepa Seshadri, Deloitte India
Next Article