The following reflection is based on inputs made by RIA Executive Director Pria Chetty at the Gordon Institute of Business Science: EMPOWERHer 2025 conference, where she participated in a panel discussion alongside Lindiwe Matlali, Naomi Thompson and Dr Amaleya Goneos-Malka on ‘Advancing Women through Technology and Innovation’.
Unmasking the digital gender gap
For years, the dominant narrative about the digital gender gap suggested that women simply did not own enough phones or have sufficient internet access. The prescribed solution was predictable: more devices, more infrastructure.
But our long-running After Access surveys, which generated some of the most granular demand-side data available across African countries, told a different story. Even when women had phones, they often lacked affordable data bundles, the necessary digital skills, or the confidence to use their devices for economically empowering activities such as e-commerce, professional networking, or job searching.
The insight was clear. The data found that access does not equal inclusion.
Our data also revealed additional layers of exclusion. Women are more likely to experience online harassment, or to have their device access monitored by men in their households. A lack of relevant local-language content and digital services tailored to women’s needs compounds the challenge. These intersecting barriers show why simply “closing the access gap” is insufficient. True digital inclusion requires addressing the structural inequities in income, education, safety, and representation that shape women’s digital lives.
Designing AI for equity
Artificial intelligence (AI) brings both risk and possibility. Left unchecked, it risks entrenching existing inequalities, reinforcing stereotypes, and amplifying harms, especially for women. But if designed intentionally, AI can be a too to embed equitable practices into workplaces.
Consider the potential of AI systems that analyse workplace interactions to identify when women’s contributions are ignored, interrupted, or uncredited. Or algorithms that review performance evaluations to expose subtle biases in language—where women are consistently praised for collaboration while men are praised for leadership. These are not speculative possibilities but feasible design choices.
This is what we mean by Equity by Design, one of nine principles that comprise RIA’s Just AI in our Framework of Inquiry. Equity by Design means embedding fairness into digital systems from the outset, so they cannot exacerbate harms in the aftermath. This could look like AI monitoring how mentorship, leadership training, or high-visibility projects are distributed, ensuring opportunities do not unconsciously flow to the same demographic. It could connect women to influential mentors and sponsors they might otherwise struggle to access, helping close informal network gaps. In this way, AI shifts from being a passive diagnostic tool to an active mechanism for fairness. However, for this to be possible, the inherent biases in AI datasets and lacking ethical standards need to be addressed.
Grounding AI in African contexts
These opportunities cannot be realised if the Global South continues to import developmental and regulatory frameworks from elsewhere. Current AI ethics debates are largely dominated by multinational corporations and scholars in the Global North. Governance models developed in those contexts rarely account for Africa’s uneven infrastructure, diverse communities, and developmental challenges. Without deliberate localisation, AI skewed in favour of the Global North risks positioning African countries, leaving its residents as mere consumers of technology rather than co-creators of innovation.
This is why RIA’s Just AI Project reframes the conversation. We propose a justice-oriented lens that goes beyond generic concepts of “responsible” or “fair” AI. Our framework foregrounds care, embodiment, equity, and restorative justice, asking how AI can actively address systemic inequities rather than reproduce them.
Embodied data and women’s lives
A critical part of this work is rethinking how we understand data itself. Too often, data is treated as an abstract resource to be mined, traded, or analysed. But as our recent volume, Embodied Data and AI: A Collection of Essays, argues, data is not disembodied—it is intimately tied to people’s lived realities, and consequently an extension of self.
One essay, Embodied Data and Deepfakes: Eroding Women’s Bodily Integrity in Digital Environments by Zara Schroeder, illustrates this vividly. Deepfake technologies that manipulate women’s images do more than violate privacy; they undermine bodily integrity itself, extending gendered violence into digital spaces.
This perspective is crucial because it underscores that AI systems, when left unchecked, can perpetuate deeply gendered harms. But in reframing our understanding of data, and by extension our understanding of digital rights, we can open the door to reimagining data practices and AI design that respects dignity, foregrounds care, and centers justice.
Towards dignity and inclusion
AI and digital technologies are not neutral; they reflect the values, priorities, and power dynamics of those who design them. To advance women through technology and innovation, justice must not be an afterthought but the foundation.
This means that we face an urgent need to not only invest in solving demand-side issues, such as device affordability and digital literacy, but also to push for developmental reform that can enforce safeguards against online harms, amplify women’s voices in design and governance, and insist on Equity by Design when modelling AI systems. It also means centring African perspectives in global debates, ensuring that the technologies shaping our futures reflect the realities and aspirations of the Global South.
Technology should not merely measure inequality—it should be harnessed to dismantle it. By grounding innovation in care, justice, and equity, we can create digital futures where women are not only present but powerful.
