Safaricom says AI prevents 94% of fraud as Kenya warns of deepfake risks

Kenya's telecommunications giant Safaricom says artificial intelligence is helping it prevent 94% of attempted fraud in real time, while government officials warn that the technology is also exposing children to sophisticated online scams, impersonation and manipulated content.

Kenyan telecommunications company Safaricom says its artificial intelligence systems can prevent 94% of attempted fraud before it occurs, demonstrating how the technology is becoming an important defence against financial crime even as cybercriminals use it to develop more sophisticated attacks.

Safaricom Chief Executive Officer Peter Ndegwa disclosed the figure during the company's cybersecurity summit in Nairobi on Wednesday, 14 October 2026, where government officials and industry representatives discussed the opportunities and risks associated with artificial intelligence.

According to Ndegwa, Safaricom uses AI to analyse customer behaviour, identify unusual transaction patterns and assess potential fraud risks in real time.

The reported 94% prevention rate is a company claim. The published account does not provide an independently audited assessment or specify the measurement period, types of fraud covered or total number of attempted incidents.

AI strengthens security across mobile money services

Safaricom's growing reliance on artificial intelligence reflects the importance of protecting digital financial transactions in Kenya and across East Africa.

The company operates M-Pesa, one of Africa's leading mobile money platforms, which processes approximately 150 million transactions daily, according to figures cited in the report.

AI-powered fraud detection systems can analyse large volumes of transactions, compare activity against established customer behaviour and flag suspicious patterns that might indicate unauthorised access or fraudulent transfers.

These capabilities allow financial service providers to respond to emerging threats more quickly than systems relying exclusively on fixed detection rules.

Ndegwa emphasised that cybersecurity is increasingly a strategic responsibility for business leaders, rather than a technical issue confined to information technology departments.

He argued that maintaining customer confidence is essential for the continued expansion of digital payments and other technology-driven services.

Safaricom also provides managed cybersecurity services to businesses, supporting the detection, monitoring and response to digital threats.

Kenyan government warns of deepfake threats to children

Despite the benefits of artificial intelligence, Kenyan authorities warned that the same technology is being exploited to create new forms of online harm.

Kenya's Cabinet Secretary for Information, Communications and the Digital Economy, William Kabogo, raised concerns about AI-generated impersonation, fraudulent messages, manipulated images and videos, and the increasing vulnerability of children online.

Deepfakes are artificially generated or manipulated audio, images or videos that can convincingly imitate real individuals or depict events that never occurred.

Such content can be used to impersonate trusted people, spread false information, facilitate financial fraud or create non-consensual intimate imagery.

Kabogo warned that children's online safety must become a central consideration in national cybersecurity strategies, particularly as young people increasingly interact with AI-powered applications, gaming platforms and digital companions.

He called for stronger cooperation between government institutions, technology companies, telecommunications operators, schools and parents to protect young internet users.

His concerns were echoed in separate reporting from the Safaricom Cyber Security Summit, which identified identity impersonation and AI-enabled fraud as growing challenges for Kenyan organisations.

Cybercrime losses highlight Africa's digital security challenges

The warnings come amid growing concerns about the financial and social consequences of cybercrime across Africa.

According to figures attributed by iAfrica to INTERPOL's African Cyberthreat Assessment Report 2026, cybercrime-related losses reported across the region increased from $192 million in 2024 to $484 million in 2025.

The report also states that AI was associated with 55% of reported cybercrimes across 36 African countries.

These figures should be understood as reported statistics from the cited assessment, rather than a comprehensive measurement of every cybercrime incident or financial loss across the continent.

The World Economic Forum's Global Cybersecurity Outlook 2026, also cited in the report, found that 87% of surveyed organisations identified AI-related vulnerabilities as the fastest-growing cyber risk during 2025.

Together, these findings illustrate the challenge facing African governments and businesses: artificial intelligence can improve threat detection while simultaneously increasing the sophistication and accessibility of cyberattacks.

Kenya seeks stronger safeguards for artificial intelligence

Kabogo said Kenya needs an approach that combines technological innovation with effective regulation, digital literacy and protection for vulnerable internet users.

The country's existing legal framework includes the Data Protection Act of 2019 and the Computer Misuse and Cybercrimes Act of 2018.

Kenya has also developed an AI Strategy for 2025–2030, while officials are considering additional policy and legislative measures to govern artificial intelligence.

According to the Cabinet Secretary, proposed measures include an Artificial Intelligence Bill, 2026, and the establishment of an Office of the Kenya AI Commissioner.

These proposals should not be confused with legislation already enacted or institutions already established.

Kabogo encouraged technology companies to improve parental controls, strengthen privacy settings and make the operation of algorithms more transparent, particularly where children are concerned.

He also emphasised the need for public education that enables families to recognise online manipulation and respond appropriately to digital threats.

Lessons for Rwanda and other African countries

The issues raised in Kenya have broader relevance for African countries expanding digital financial services and adopting artificial intelligence.

In Rwanda, the National Cyber Security Authority has identified AI-generated impersonation, phishing, fraudulent calls and other forms of digital deception among the threats addressed by its 2026 cybersecurity awareness campaign.

The authority recommends independently verifying suspicious communications, using multi-factor authentication and reporting suspected cyber incidents through official channels.

These measures complement the wider emphasis on institutional accountability, public awareness and secure technology deployment discussed at the Nairobi summit.

For Safaricom, the immediate priority remains strengthening fraud prevention while maintaining trust in its expanding digital services.

For Kenyan authorities, the challenge extends beyond protecting networks and financial transactions to ensuring that children and other vulnerable internet users are protected as AI becomes more widely available.