Sam Altman. Elon Musk. Two of the recurring names in the realm of AI. These names are on the lips of a wide range of news outlets and conferences making it difficult to talk about AI without mentioning their names. Currently, the world is up in arms about the cat-and-mouse game that OpenAI is playing with Sam Altman. He was kicked out of his own company, scouted by Microsoft and is referred to as the unofficial godfather of AI. Altman is placed on the pedestal because he believes that AI will make a better world for ‘us’. His vision is probably why Microsoft wasted no time in making him theirs. However, the ‘us’ in question warrants some investigation. Who is ‘us’ and why does it appear that in the narrative of AI, women have not and—perhaps—will not exist?
In an Altman-led AI universe, do women not exist?
The furore that envelops Altman being chucked by his own company is an interesting turf to venture into. But it also leads to a considerable number of revelations. As mentioned in The Guardian, the current composition of the Altman team reveals a shocking absence of women, sparking considerable attention and prompting an analysis of the situation surrounding his layoff. A closer look at the controversy unveils intriguing insights; dishearteningly, over 75% of the 702 employees advocating for Altman’s reinstatement were men. This gender disparity aligns with findings from McKinsey’s The State of AI in 2022 report, emphasising the prevalent gender imbalance in AI teams.
Altman’s reinstatement brought about a newly formed board of directors at OpenAI, comprising exclusively of white men. This lack of diversity extends to executive roles, perpetuating a male-dominated hierarchy. The glaring question arises: where are the voices of female leaders and experts in the narrative of this Silicon Valley saga? The marginal presence of women in the AI industry raises concerns, with their perspectives likely going unnoticed and unaddressed. This absence becomes particularly apparent when examining women’s roles in shaping our AI-driven future and influencing the discourse surrounding generative AI. A deep dive into data and expert discussions underscores the stark reality: women, whether as developers, news editors, or AI experts, are conspicuously underrepresented in the AI landscape.
Generative AI (GAI), reliant on vast datasets of text, images, and video, has historically featured a predominantly male representation. This ingrained bias, mirrored in news coverage, coupled with the existing structural gaps faced by women in society, culminates in a narrative about GAI’s risks, limitations, opportunities, and trajectory predominantly shaped by male perspectives.
Where are the women behind the scenes?
It all boils down to a matter of who is seated at the head of the table and who runs the conversation. And it is not a secret that it is men who have the upper hand in this situation. Women being underrepresented in AI is a grave issue. One that needs to be solved immediately. The absence of diverse perspectives in technology development poses a significant risk, as it may lead to the creation of new technologies that inadequately address the needs of a substantial portion of the population. As explained in an online article posted by Pursuit-University of Melbourne, this concern is underscored by the European Commission’s 2020 white paper on AI, which advocates for the implementation of measures to prevent AI systems from resulting in outcomes that involve prohibited discrimination.
Moreover, there is a risk of overlooking innovations in activities predominantly undertaken by women. The impact of AI on gender biases is widely acknowledged, as AI systems inherently mirror the biases present in the data they are trained on. When data contains gender stereotypes, AI perpetuates these biases.
This imbalance also persists in the research available about women who work in AI. There is a serious limitation of intersectional data in the tech sector. As mentioned in a research article titled “Feminism Confronts AI: The Gender Relations of Digitalisation” by Judy Wajcman and Erin Young (published on Oxford Academic), only 1.6 percent of Google’s US workforce in 2020 comprised Black women, highlighting a stark disparity. Despite the implementation of diversity policies, training, and various initiatives, the impact of increasing the presence of women of colour in tech remains marginal, as elucidated in Google’s 2020 data.
Globally, the World Economic Forum (2020) estimates that women constitute approximately 26 percent of workers in data and AI roles, a figure that drops to 22 percent in the UK. This disparity is attributed in part to the ambiguity and novelty of these professions, but it is also exacerbated by a reluctance among major tech companies to disclose this data. The current data on gender diversity in the AI field is disheartening, characterised by partial and often flawed information released by AI companies. This lack of transparency poses a significant obstacle to comprehensive research in this area.
Furthermore, it is important to understand that the operation and longevity of ‘intelligent’ machines are heavily dependent on an extensive and often unnoticed human labour force. This workforce performs crucial technological tasks such as data labelling for algorithmic inputs, code cleaning, machine learning tool training, and content moderation and transcription. These individuals, commonly referred to as ‘ghost workers,’ are frequently women residing in the Global South. Unfortunately, they face issues of underpayment, undervaluation, and a lack of protective labour laws. The exploitation and marginalisation of these ghost workers happen strongly and predominantly within the female workforce.
Sadly, this is nothing new. Women have been and continue to be missing from the stories that are told about the tech industry. It is very rare that the general public is made aware of the contribution that women have made to propel the developments in the field of AI. To add fuel to the fire, tech companies are hesitant to hire women to be a part of their workforce. The irony of the digital era democratising the world is that women still have a long way to go before they are emancipated from the clutches of a patriarchal system, one that is set in stone even in the field of AI.
(Sandunlekha Ekanayake)