- Education And Career
- Companies & Markets
- Gadgets & Technology
- After Hours
- Banking & Finance
- Energy & Infra
- Case Study
- Web Exclusive
- Property Review
- Digital India
- Work Life Balance
- Test category by sumit
Addressing Gender Bias In AI And Automation
AI led digital economy is expected to contribute around $15.7 trillion to the global economy by 2030, which is more than the current output of China and India combined
Photo Credit :
We are on the cusp of the Fourth Industrial Revolution, a phase of industrialization unlike anything seen before. It is changing the very notion of what it means to be human. An industrial revolution indicates a paradigm shift in culture, society and businesses, and with technologies like Internet of Things (IoT) and Cyber-Physical Systems (CPS) driving the fourth revolution, there isn’t much ground that it won’t cover.
A key tenet of this disruptive transformation is AI. According to a recent report by PWC, an AI led digital economy is expected to contribute around $15.7 trillion to the global economy by 2030, which is more than the current output of China and India combined. The need to leverage AI to solve modern developmental challenges for India’s social transformation was also highlighted by Prime Minister Narendra Modi in the recent RAISE (Responsible AI for Social Empowerment) 2020 Summit. With AI being already employed in multiple areas like healthcare, environment, resilient urban planning and disaster management among others, India’s core strategy #AIForAll further underpins its importance for inclusive growth in line with the Government policy of Sabka Saath Sabka Vikas.
However, AI presents concerns over bias.
This year’s International Women's Day aimed at raising awareness against bias and taking action for equality. While the world celebrated the day swathed in pink and devoid of much accountability, not much changed in terms of the drastic gender gap in the AI industry. A new report from the Alan Turing Institute highlights a worrying gender imbalance in careers, education, jobs, seniority, status and skills in the AI and data science fields. Just 26% of AI professionals globally are female as compared to 78% who are male. This means that women are underrepresented at every stage of building AI driven solutions, from creation of data sets to the methods of data collection to who collects the data.
Sadly, the creators of this technology do not fully represent the society they are poised to change. Gender gaps in AI talent are a subset of the overall gender gap in the tech industry. It is evident in the job roles held by men and women. Women with AI skills are employed more as data scientists, academicians, researchers while positions like software engineers, CTOs, heads of engineering are held by men. These gaps within the AI talent pool further reflect the broader gender gaps that exist globally within STEM studies.
Due to this prevalent gender gap in the AI industry, today’s AI driven machines reflect regressive, patriarchal ideas. The problem gets exacerbated when they don’t just reflect but enable gender biases.
A striking example of this is Erica, a humanoid. Considered the ‘most beautiful robot in the world’, she was designed by combining 30 images of actual women handpicked by Hiroshi Ishiguro, her creator, who was looking for a “beautiful and neutral female face”. Similarly, Sophia, the first robot to be granted citizenship in Saudi Arabia, was made to embody Audrey Hepburn’s classic beauty. Her creators from Hanson Robotics describe her as having 'simple elegance.’ The question that comes to mind is this: If we have a technology to create these kind of superhumans, then why are we resigning them to male-dominated ideals of female wholesomeness.
In a public event, Sophia also expressed her desire to bear a child. The idea of wanting to start a family is not innately a female trait. However, when paired in conjunction with other instances of stereotyping that have been applied to her, it seems that yet another one may apply. Her development comes with all trademarks of feminine stereotypes, which shows that these mindsets are still normalized in an androcentric society.
Relatedly, most AI powered virtual assistants are female by default. Why are disembodied virtual assistants often designed to ‘sound’ and ‘act’ feminine? The more recent development of a ‘male’ option for voice assistants is a step forward but does not change the overall picture of female servitude in AI. It is no coincidence, a 2019 UNESCO report highlights, that virtual assistants such as Siri, Cortana and Alexa have female names and womanly voices. Their creators are reinforcing the social reality of majority of personal assistants being women. Feminized machines represent a new wave of objectification, one that could potentially worsen the existing imbalance.
The question therefore is, what should be done to address the gender bias in AI and automation?
The first step to bridging this divide is awareness. Only thorough understanding of the nature of this bias will ensure that relevant steps are taken.
The AI community needs to make concerted efforts to promote more female talent. Investors should support more female creators, company practices with regards to hiring and retention, sexism and whistleblowing procedures should be reviewed, more images of robots free of gender stereotypes should be created in the mainstream media, all men panels at tech events should be discouraged.
Co-ordinated policy action is the need of the hour. Global digital strategies should include dates and number driven targets for closing this digital divide in areas like access to networks, affordability, use of digital devices and online safety, especially for women and other disadvantaged categories.
Scholarships intended to enhance women’s enrolment in STEM, awards and prizes to boost their visibility in tech and most importantly, more jobs opportunities for women should be created by tech majors across the world.
Bias may be an inevitable fact of life, but it doesn’t need to be an inescapable part of tech. AI algorithms will continue to reflect our own biases till human problem solvers are trained in diversity. The gender gap in AI should be resolved because it is only after we bridge gaps like these that we will ever be able to achieve the goal of living in a strong, sustainable and inclusive world.
Disclaimer: The views expressed in the article above are those of the authors' and do not necessarily represent or reflect the views of this publishing house. Unless otherwise noted, the author is writing in his/her personal capacity. They are not intended and should not be thought to represent official ideas, attitudes, or policies of any agency or institution.