A constitutional petition filed on April 2 at the High Court in Nairobi challenges how Safaricom, Kenya’s largest telco, deploys artificial intelligence (AI) across its telecoms and mobile money services, raising questions about automated decision-making in systems used by millions of Kenyans.
The case names Safaricom, the Attorney General, the Communications Authority of Kenya (CA), the Competition Authority of Kenya (CAK), and the Consumer Protection Advisory Committee as respondents.
The filing seeks court orders that would require Safaricom to disclose how its algorithms operate, introduce human review in key processes, and submit its systems to regulatory oversight. It also asks regulators to enforce safeguards around AI use in essential services that underpin daily transactions and access to credit.
The petitioner, Victor Odhiambo, states that Safaricom uses AI to process large volumes of user data and automate decisions across mobile money, credit scoring, fraud detection, and account management. These systems can influence whether transactions are approved, accounts are restricted or users can access short-term credit products such as overdrafts.
“The extraction of consumers’ personal data… and their behavioural patterns is not incidental to the delivery of telecommunications and financial services but is itself a primary commercial activity,” the petition states.
The document points to Safaricom’s customer service structure, including its Zuri chatbot, which it says handles large volumes of queries. Safaricom says about 85% of daily customer interactions are handled through automated channels, including phone-based voice menus, USSD prompts and the Zuri chatbot.
The petition further argues that users often interact with automated systems when seeking support and may not receive clear explanations or access to human review when disputes arise or transactions are contested.
The filing says users are not given sufficient information on how decisions are made, what data is used, or how to challenge outcomes such as transaction blocks, reversals, or credit limits. It cites concerns over large-scale data collection, including behavioural and biometric information, and the absence of publicly available impact assessments.
“Algorithmic decisions that block funds, restrict accounts, deny reversals… without intelligible reasons, notice, and meaningful human review” fall short of constitutional standards, the petitioner argues.
The case links these claims to constitutional rights, including privacy, consumer protection, equality, and fair administrative action. It argues that systems affecting access to essential services must meet those standards, even when operated by private companies with significant market reach.
The petition also states that Kenyan authorities have identified risks linked to artificial intelligence but have not established clear enforcement mechanisms as deployment expands.
The remedies sought include orders requiring Safaricom to publish information on its automated systems, establish human oversight mechanisms, and conduct impact assessments. The petitioner is also asking the court to direct regulators to take action and to retain supervisory jurisdiction over compliance.