When I joined Stanford’s Tech Ethics & Public Policy Practitioner Program, I carried with me the realities of the Global South — my own reality as an African Arab woman living with a chronic illness, coming from a society where digital transformation is both a promise and a paradox. In the North African context, artificial intelligence unfolds amid fragile institutions, deep data inequalities, limited internet access, and the high cost of computing education.
Stanford gave me analytical tools — a harmony of political, economic, social, and historical perspectives that revealed how technology is never isolated from power. I learned how Silicon Valley emerged, how the foundations of the Fourth Industrial Revolution were laid, how the internet reshaped global hierarchies, and how large language models now influence even the political calculus of the White House. Yet the deeper lesson was that AI is not neutral. For Africa, this realization means refusing imported frameworks and instead grounding AI governance in Ubuntu and Hikma — philosophies of interdependence and wisdom that restore humanity to technology. To lead with values, as we explored at Stanford, is not to decorate policy with principles; it is to redistribute power — to make decisions through the eyes of the vulnerable: workers, job seekers, women, and those living with chronic illness and caregiving responsibilities. It is to ask whose lives are seen, whose voices are heard, and whose suffering is hidden within the efficiency of code.
- Seeing the Lives Algorithms Control: Moving from Technical Fixes to Systemic Justice
Toni Morrison reminds us that before solving, we must dream — a moral imagination that enters another’s world. Without that empathy, technical solutions often reproduce the very injustices they claim to fix. When we speak of “AI transformation” or “digital upskilling,” we rarely ask what the worker feels when an algorithm cuts her pay. Morrison called this the preamble to problem-solving — we cannot fix what we have not yet dared to feel.
Across Africa, millions already live under algorithmic systems built without moral imagination. Mohamed, a driver in Tunis, works for three platforms because none provides enough income. When a false complaint lowered his rating, the algorithm cut his rides by 40 percent — no human review, no appeal. Aminatou , a data labeler in Dakar , earns less than a dollar per thousand images, moderating violent content without psychological support. Fatima, a textile worker in Morocco and a cancer survivor, is monitored by cameras that track her every move. Her pay is docked whenever she pauses to rest during the hot flashes caused by her ongoing hormonal treatment.
Their stories reveal automation without accountability. The UNESCO Recommendation on the Ethics of AI (2021) calls for human oversight — systems where people can intervene, review, and protect dignity. Yet in Africa , neither human-in-the-loop nor human-on-the-loop mechanisms exist. Mohamed needs review before pay cuts, Amira needs mental-health protection, and Fatima needs a union empowered to challenge algorithmic penalties. UNESCO’s principle is clear: AI can never replace human responsibility. But today, no one is accountable when their livelihoods are lost.
- Beyond Fairness: Why Algorithmic Justice Is a Category Error:
Computer scientist Arvind Narayanan argues that focusing on “algorithmic fairness” — making algorithms statistically unbiased — misses the point entirely. When Mohamed’s income was cut, was the problem bias? Even if the algorithm treated all drivers equally, he would still lack recourse, union protection, or recognition as a worker. When Amira labels trauma for $0.80, even “equal pay” would not grant her mental health care or labor rights. When Fatima’s score drops for helping others, the issue isn’t bias — it’s that the system treats solidarity as inefficiency. Fairness is a bandage for a bandage. As Narayanan notes, when algorithmic decision-making itself is a symptom of deeper structural failure, focusing on fairness is “two levels removed from what needs to be fixed.”
Disparities are symptoms, not causes. Platform work exists because Tunisia’s labor market offers few formal opportunities. Data labeling thrives because graduates have no jobs. Surveillance spreads because buyers demand it and regulators stay silent. Moreover, people with disabilities and chronic illnesses are already marginalized in the workforce — often excluded from stable employment and left without protection. Under today’s “independent contractor” model, employers hold no obligations, and unions cannot intervene to defend these
Beneath these systems lies what King and Meinhardt (2024) describe in Rethinking Privacy in the AI Era: Policy Provocations for a Data-Centric World (Stanford HAI) as the transparency illusion — the myth that users understand or control how their data is used. AI’s hidden infrastructure is a data supply chain stretching from users who click “accept all” to workers who clean, label, and moderate that data. Companies disclose their practices through long, unreadable terms of service, creating the appearance of consent while concealing deep asymmetries of power.
In the Global South, this illusion is reinforced by a cultural belief in technological knowledge and wisdom — the assumption that technology is objective, inevitable, or even benevolent. This discourages scrutiny and allows data flows to reproduce inequality, labeling, gender-based violence, and exclusion. Such blind trust in emerging technologies can increase the vulnerability of people living with chronic diseases or mental health trauma.
A subtler harm emerges in AI intimacy — when users confide their emotions to chatbots or virtual assistants. These exchanges, often born of loneliness or illness, become another layer of monetized data. As a breast cancer patient, I often ask questions about fatigue or supplements — unaware that my words might train health-marketing algorithms. My private conversations with a chatbot could refine predictive systems designed to target my consumption patterns.
Without ethical design and moral responsibility, this becomes the conversion of vulnerability into profit — without consent.
To understand why such contradictions persist, we must examine the moral economy of those shaping technology. As Neil Malhotra and David Broockman observe, Silicon Valley’s elite embodies a “liberal-tarian” ethos — liberal in values but libertarian in structure. They champion inclusion, diversity, and climate action rhetorically, yet resist regulation and labor oversight that might constrain market autonomy. This duality explains why the tech industry can appear morally progressive while remaining politically resistant to accountability.
In theory, these leaders uphold integrity, responsibility, fairness, and stewardship — the very values emphasized in Leading with Values. Yet each is applied selectively. Integrity shows up through public advocacy for openness and equality but collapses the moment oversight threatens autonomy.
Conclusion
To end, Stanford’s Tech Ethics & Public Policy Practitioner Program taught me the value of a philosophical approach in the era of technology and modernization. Morality and humanist angles are essential to finding social solutions to the harmful impacts of AI — especially in regions still lacking strong governance frameworks. The Global South remains deeply vulnerable to digital colonialism, AI-driven harm, and rising unemployment, as the workforce is not yet ready to engage with or fully understand automated systems.
References
King, J., & Meinhardt, C. (2024). Rethinking Privacy in the AI Era: Policy Provocations for a Data-Centric World. Stanford Institute for Human-Centered Artificial Intelligence (HAI).
UNESCO. (2021). Recommendation on the Ethics of Artificial Intelligence.
Narayanan, A. (2018). Translation Tutorial on Algorithmic Fairness. Princeton University.
Malhotra, N., & Broockman, D. (2017). Liberal-Tarian Values and the Tech Elite. Stanford Graduate School of Business.
