Kenya’s laws fail to protect women from AI-enabled abuse

Kenya’s laws fail to protect women from AI-enabled abuse


By Faith Muiruri


Artificial intelligence (AI) is transforming how Kenyans communicate, learn, work and


participate in public life. But the rapid spread of AI tools is also opening new avenues for


gender-based violence that the country’s legal and institutional systems are not yet prepared


for.


A new analysis by the Community Advocacy and Awareness Trust (CRAWN Trust) says Kenya


faces a widening regulatory vacuum as deepfakes, voice cloning, synthetic intimate images,


automated harassment and AI-generated disinformation become easier to produce and


distribute.


The study, funded by the African Women’s Development Fund, describes technology-facilitated


gender-based violence as a rapidly evolving threat that affects not only personal safety, but also


elections, education, financial inclusion, and national development.


Among the most urgent risks, is the creation of non-consensual intimate imagery using


generative AI. “Such images can be fabricated without any physical encounter between the


perpetrator and victim. A photograph posted online may be manipulated into sexually explicit


material within minutes and distributed across multiple platforms before the victim even knows


it exists,” says the report.


The report further notes that the damage is often immediate and difficult to reverse. “Harmful


content can be copied, reposted and stored indefinitely.”


It refers to this as a “digital tattoo” where a permanent record of abuse, forces survivors to


relive the violation whenever the material reappears.


Currently, Kenyan laws do not specifically criminalise the creation of AI-generated non-


consensual sexual imagery. The study observes that existing provisions may address some


forms of harassment or the distribution of intimate images, but they do not adequately cover


synthetic content, voice cloning, or the tools used to create it.


Electioneering period


“This legal uncertainty becomes even more dangerous during an electioneering period,” adds


the report.


What is worrying is that AI-generated videos, audio and images could be used to impersonate


women candidates, spread sexualised rumours, distort their statements or create false


evidence.


“With Kenya’s high mobile penetration and uneven digital literacy, fabricated content could


circulate widely before it is verified.”





The report explains how coordinated online campaigns can combine fake accounts, doxing,


abusive comments and misleading narratives to damage a candidate’s reputation and


discourage her from campaigning. “Such attacks may appear to be ordinary political criticism,


but the report argues that many are designed to silence women and exclude them from public


participation.”


The risk extends to journalists and activists. “AI-generated audio could be used to falsely


attribute statements to a reporter or human rights defender. Synthetic videos could undermine


public trust in legitimate investigations. Women who speak about corruption, governance or


violence may become targets of coordinated disinformation intended to force them offline.”


Artificial Intelligence Bill


The report raises concerns about the Artificial Intelligence Bill, 2026, currently before the


Senate. It describes the Bill as an important starting point, but says it requires substantial


amendment before enactment.


The report singles out institutional fragmentation as one of the areas that need further


consideration. According to the report, the proposed creation of several new AI oversight


bodies alongside existing institutions such as the Office of the Data Protection Commissioner


and the Communications Authority, could leave survivors unsure where to report.


“A woman whose likeness has been used in an AI-generated sexual image might have to deal


separately with police investigators, data regulators, cybercrime bodies, and online platforms.”


The report recommends that the AI Commissioner, if established, should coordinate with


existing institutions rather than create parallel and competing systems. It calls for a single,


integrated complaint pathway for AI-related gender-based violence.


The Bill should also classify systems designed to generate non-consensual sexual imagery or


impersonate real people for harmful purposes as prohibited-risk applications. It should include


specific protections for children, electoral safeguards and mandatory testing against cyber


threats such as model poisoning and prompt injection.


For children, technology-facilitated abuse often begins in schools, peer messaging groups,


online classrooms and social media. “Girls may face cyberbullying, grooming and the circulation


of intimate images, while boys may be exposed to online communities that normalise misogyny


and entitlement.”


Of concerns to children rights advocates is that the rollout of AI-enabled learning tools under


the Competency Based Curriculum has also raised questions about children’s data. The report


says parents may not have sufficient information about how behavioural, biometric, or learning


data is collected, processed and used.


Children Act





It recommends amending the Children Act to recognise technology-facilitated child abuse


explicitly. “Digital platforms used by children should have age-appropriate safeguards,


accessible reporting tools and parental consent mechanisms.”


“The Ministry of Education should oversee AI systems used in schools in coordination with data


protection and cybercrime authorities.”


In addition to children, the report notes that women are experiencing serious economic digital


abuse. Control over M-Pesa credentials, forced disclosure of passwords, monitoring of


transactions and unauthorized digital loans are being can be used to trap women in abusive


relationships. “Yet financial regulations do not clearly recognise digital financial coercion as


gender-based violence.”


The report calls on the Central Bank of Kenya to establish reporting and remedy pathways for


technology-enabled economic abuse. Mobile money and digital finance providers should


develop safety triggers for suspicious transaction patterns and mechanisms to help survivors


restore control over their accounts.


Digital protection orders


This should be treated with urgency since protection from service providers remains an area of


concern. Google transparency report, for instance, shows that approximately 62 per cent of


Kenyan government content-removal requests were rejected during the first half of 2025. For


survivors of intimate-image abuse, delays can mean that content is viewed and copied


thousands of times.


To ensure compliance by such companies, the proposed reforms include requiring platforms to


act within 24 hours to remove offending content, stronger penalties for non-compliance, and


mandatory local content moderation and transparency reports on complaints. The study


recommends online platforms to develop moderation capacity in Swahili and other local


contexts.


Other recommendations in the report include: Development of county-level forensic capacity


and specialized prosecution units; creation of precise, proportionate laws that distinguish


lawful speech from threats, harassment, intimidation, sexual exploitation and coordinated


digital violence; and ratification of the International Labour Organization’s Convention 190,


which recognises violence and harassment in technology-enabled workplaces.


Kenya’s AI future is being shaped today. If safeguards are delayed, the country may find itself


responding to harms that have already spread beyond control. If reform is pursued urgently


and coherently, technology can remain a tool of innovation rather than becoming an


instrument of intimidation.





--SS--