AI and the future of SGBV: Can we stop the next wave of violence?

By  Moraa Obiria

Senior Gender Journalist

What you need to know:

• Manipulated images, cloned voices and deepfakes are creating new avenues for technology-facilitated sexual and gender-based violence against women online.
• Experts warn that AI can amplify harassment, hate speech, impersonation and sexual abuse while calling for stronger platform accountability and online safety.

After Soipan Tuya was reassigned to the Ministry of Defence as Cabinet Secretary (CS) on July 23, 2024, following President William Ruto’s reconstitution of his Cabinet, she became a target of online abuse.

Soipan, who had previously served as CS for Environment, Climate Change and Forestry, was subjected to manipulated content circulating online. An artificial intelligence (AI)-manipulated photo showing her and President Ruto embracing was widely shared on X.

A manipulated video also circulated on TikTok, with President Ruto's voice edited to make it appear as though he was saying Soipan used to be his girlfriend.

Fact-checked

Africa Check fact-checked the fake items and established that the photo was from a 2013 event, based on comparisons with photographs from the period posted on President Ruto’s Facebook page.

The original photographs showed the two walking and sitting apart, contrary to the manipulated image, which made them appear to be in a close embrace.

It also traced the video to the original footage posted on State House’s YouTube account. The original recording revealed that President Ruto’s “girlfriend” remark referred to his wife, Rachel Ruto, not Soipan.

The case illustrates how technology can be used to create and spread false and sexually suggestive content targeting women, an emerging form of technology-facilitated sexual and gender-based violence (SGBV).

The United Nations says AI is accelerating existing forms of technology-facilitated SGBV, including sexual harassment, stalking, hate speech, misinformation, defamation and impersonation. 

It is also enabling new forms of abuse, including hacking, astroturfing, video- and image-based abuse such as deepfakes, doxing, cyberbullying and online grooming.

Documented cases

In Kenya, the Kenya ICT Action Network documented 58 cases of online gender-based violence on its tracker between July and September 2026. Women and girls were targets in 43 of the cases.

The reported cases included cyber harassment and bullying (29), gender-based hate speech (26), online sexual exploitation (12), digital voyeurism and cyberflashing (six), and non-consensual sharing of intimate images (four).

Dr Erick Omuya, an AI researcher and chairperson of the Department of Computing and Information Technology at Machakos University, says AI has increased the potential for gender-based violence in the digital space.

“Before the introduction of AI and other digital technologies, most of this happened manually. We had people being attacked, experiencing sexual violence and all that,” he says.

But generative AI has created new avenues for abuse, he says, allowing perpetrators to create and distribute fabricated sexual content.

“Instead of people taking photos and sharing, they can use generative AI to generate pornographic material, nudes and share,” Dr Omuya says.

Cloning voices

He says perpetrators can also take a person’s image and use AI to create false material about them, which can then be used to harass or discredit them.

“You can clone voices using AI and say something gender-related and attribute it to somebody. It could be somebody of high standing in society, maybe like a politician,” he says.

AI can also facilitate coordinated digital attacks, with bots used to issue threats and spread hate speech, Dr Omuya says. Their anonymity can make perpetrators difficult to identify, he says.

He adds that impersonation, disinformation and misinformation are other ways AI can be used to perpetrate gender-based violence.

But he says the same technology can help counter these threats. Experts, he says, can identify and block bots used in coordinated attacks, much as firewalls filter devices and information on networks. 

For people without technical expertise, he says, AI tools can also guide responding to online abuse, accessing support and protecting their mental health.

“There are initiatives that use some kind of support for both survivors and individuals who may feel targeted and so they want to protect themselves,” he says.

Cautious on information shared

Dr Omuya, however, says individuals also need to be cautious about the information they share online.

“People can contribute to or attract abusers by how they use their devices,” he says. “Those who are highly visible online may make it easier for potential abusers to track their activities. Posting photographs and videos on platforms such as WhatsApp and Instagram from public accounts can also expose personal information that could later be used against them.”

He advises users to practise basic online safety by sharing only essential information and making their social media accounts private to limit access to personal details. However, Dr Omuya says most platforms currently lack effective safeguards to prevent publicly shared content from being copied, duplicated or redistributed.

Accountability

He says accountability starts with platform design, with developers considering the potential risks to their target users from the outset. He says platform owners and developers have a responsibility when their services are used to perpetrate gender-based violence, while governments have a role in ensuring platforms comply with ethical standards and relevant policies.

Kenya can also consider restricting access to platforms where necessary, Dr Omuya says, citing China, where some applications are inaccessible. 

But he notes that such restrictions would need to take into account the use of virtual private networks, which can circumvent them. He says the ethical development and use of AI is becoming increasingly important for developers and researchers.

“AI as a technology is supposed to help us do things more efficiently, but we also need to consider the moral fabric of individuals and the country as we use it,” he says.