Safaricom is facing a landmark legal challenge that could redefine the relationship between African customers and Artificial Intelligence.
A lawsuit filed in the High Court by Benedict Kabugi Ndung’u, following a series of petitions concerning data privacy, now targets the company’s black box automation. The suit argues that Safaricom’s increasing reliance on AI for M-Pesa credit scoring and customer service is not just a technological shift, but a violation of constitutional rights.
At the center of the controversy is the perceived dehumanization of essential services. For millions of Kenyans, M-Pesa is not a luxury; it is a lifeline. Yet, decisions regarding credit limits on products like Fuliza and M-Shwari are now largely dictated by algorithms that analyze transaction patterns, airtime usage, and behavioural data.
The petitioner alleges that Safaricom has systematically replaced human judgment with automated systems, like the chatbot Zuri, while making it increasingly difficult for customers to reach a human agent. This automation wall has left users trapped in loops with bots that lack the empathy or authority to resolve complex financial grievances, a move the lawsuit describes as an abdication of corporate responsibility.
The UNESCO Standard
The legal challenge arrives at a time when global standards are shifting toward Human-Centric AI. The UNESCO Recommendation on the Ethics of Artificial Intelligence, adopted by 193 member states including Kenya, provides a scathing critique of Safaricom’s current trajectory.
According to UNESCO’s framework, Human Oversight and Determination is a core ethical principle. The recommendation explicitly states:
“Member States should ensure that AI systems do not displace ultimate human responsibility and accountability. Persons should always have the opportunity to interact with a human rather than an AI system for significant decisions.”
By forcing users to interact solely with Zuri or automated credit-limiters, critics argue Safaricom is ignoring the UNESCO mandate that users must have a clear right to human intervention. When an algorithm denies a micro-loan that a family needs for school fees or medical bills, the lack of a human-led appeals process creates what the lawsuit calls a vacuum of accountability.
Beyond the lack of human contact, the lawsuit targets the explainability of Safaricom’s AI. UNESCO’s ethical guidelines demand that “people should be made aware when a decision is informed by AI” and that the logic behind such decisions must be transparent.
The petition contends that Safaricom’s algorithms operate as a black box. Users are often left wondering why their credit limits have dropped or why their accounts have been flagged, with no way to see the underlying data or the logic used by the machine to profile them. This, the suit argues, violates Article 31 of the Kenyan Constitution (the right to privacy) and the Data Protection Act, which requires informed and explicit consent for data-driven profiling.
The petitioner is seeking a forensic audit of Safaricom’s AI systems and a temporary injunction on automated decision-making until a proper regulatory framework is established.