Kenya’s leading telecommunications firm, Safaricom, is facing a landmark lawsuit at the High Court, challenging its widespread use of artificial intelligence across key services, including M-PESA and its Zuri chatbot.
Filed on April 21, 2026, the case argues that the company’s reliance on automated systems in financial decision-making and customer service could violate constitutional rights and data protection laws.
The suit, led by petitioner Victor Odhiambo alongside other plaintiffs, raises several critical concerns about how Safaricom deploys AI.
The lawsuit claims that Safaricom uses opaque AI systems to make significant financial decisions affecting millions of users. These include setting credit limits for overdraft and loan services such as Fuliza and M-Shwari.
Petitioners argue that users are not given clear reasons when credit limits are reduced, loan requests are denied and transactions are flagged
This lack of transparency, they claim, turns the system into a “black box” where decisions are made without accountability.
According to the petition, customers calling support lines such as 100 or 234 are frequently routed through automated systems (including the Zuri chatbot) without a guaranteed path to a human agent.
This becomes especially problematic in urgent situations like suspected fraud or blocked accounts.
The case leans heavily on provisions in Kenya’s Data Protection Act 2019, which grants individuals the right not to be subjected solely to automated decision-making and to receive meaningful explanations for decisions that significantly affect them
Plaintiffs argue that Safaricom fails to meet these requirements, particularly when users are not told why financial decisions have been made.
Another key issue raised is the risk of algorithmic bias.
The petition claims that AI models relying on generalized behavioral data may not adequately account for individual circumstances.
As a result, the system could unfairly disadvantage certain users while prioritizing what the suit describes as “algorithmic commercial optimisation” over consumer rights.
In its preliminary response, Safaricom has pushed back against the claims, defending its use of AI as both necessary and beneficial.
The company argues that its systems are essential to maintaining the stability of its platforms, noting that M-PESA handles over 2,000 transactions per second.
Without automation, it says, such a scale would be difficult to manage efficiently.
Safaricom has also raised concerns about being required to disclose how its algorithms work.
According to the telco, revealing the specific variables behind its AI decisions could allow bad actors to manipulate the system, potentially leading to increased loan defaults, fraud, and broader platform instability.
The High Court has already certified the matter as urgent, underlining its significance.
