Kenya has signed a cooperation framework with U.S. artificial intelligence company Anthropic covering AI skills, research, public-sector applications and safety, but the agreement raises questions about how a country that remains heavily dependent on foreign AI infrastructure will govern technologies that can influence elections, process sensitive data and potentially pose catastrophic risks.
The Joint Declaration was signed in New York on the sidelines of the 81st United Nations General Assembly by Foreign Affairs Principal Secretary Abraham Korir Sing’Oei and Anthropic’s Head of Beneficial Deployments, Elizabeth Kelly.
Kenya’s Special Envoy on Technology Philip Thigo and Anthropic Global Impact Lead Ambassador Courtney O’Donnell witnessed the signing.
Thigo said the partnership was intended to build Kenya’s own capacity to develop, assess and govern advanced AI rather than simply adopting technologies developed elsewhere.
“This partnership will empower Kenya to build local AI expertise, develop responsible technologies tailored to our needs, and ensure that advances in AI strengthen our institutions and drive inclusive growth,” he said.
The agreement covers AI skills and capacity building, research, responsible public-sector applications, AI safety and evaluation, and possible applications in education and healthcare. Future programmes involving access to or processing of data will be subject to Kenyan law, including data protection, cybersecurity and public-sector data requirements.
But the declaration does not itself establish a specific nationwide AI deployment. It provides an umbrella framework for future pilots, research, institutional partnerships and training programmes.
That distinction is critical as Kenya’s AI ambitions remain constrained by infrastructure, skills and public-sector capacity. Oxford Insights’ 2024 Government AI Readiness Index ranked Kenya eighth in Africa and 93rd globally, with an overall score of 43.56. The index assesses government readiness, technology-sector capacity and infrastructure.
Oxford Insights’ 2025 assessment said sub-Saharan Africa continued to face significant gaps in AI infrastructure and public-sector adoption, although Kenya had developed a promising innovation ecosystem and ranked 22nd globally for AI resilience.
The infrastructure gap is broader still. Africa has about 18% of the world’s population but less than 1% of global data-centre capacity, according to Boston Consulting Group. Less than 2% of the continent’s roughly 2,000 languages are supported by large language models, BCG said.
That leaves Kenya in a difficult position. It wants to develop sovereign capacity in AI while relying on companies whose models, computing infrastructure and safety systems are largely developed outside the country.
In the deal, Kenya commands very little control if the technology becomes central to government services, political communication or national infrastructure.
The political risks are already visible. In a September report, Anthropic said it had identified a Kenyan operation that used Claude to mass-produce political content designed to appear to come from ordinary citizens ahead of the 2027 general election. The company said the operator generated batches of exactly 50 posts, with Claude instructed to make them look like spontaneous grassroots commentary.
Anthropic said the operation promoted Energy Cabinet Secretary Opiyo Wandayi and pushed claims that the United Opposition coalition was breaking apart, targeting opposition figures including former Deputy President Rigathi Gachagua and former President Uhuru Kenyatta.
The company said it found no evidence of government involvement and described the operation as a local Kenyan political astroturfing campaign. It also said the operation failed to reach or influence real people, although the case demonstrated how generative AI could be incorporated into organised political influence efforts.
The significance for Kenya goes beyond the particular operation. Political manipulation through technology is not new in the country, where social media, coordinated online messaging and paid influence networks have become increasingly important to electoral communication. Anthropic’s findings show that generative AI can reduce the cost of producing such material while increasing its volume and making coordinated activity appear organic. Kenya’s editors have also warned that AI and disinformation could make the information environment more difficult to navigate ahead of the 2027 election.
Anthropic itself has acknowledged that AI could give authoritarian governments new tools for surveillance and repression and could be used to flood public life with misinformation. The company says governments, rather than AI companies alone, should establish rules for powerful AI systems.
That creates a critical tension in Kenya’s new partnership. Anthropic is simultaneously a technology provider, a developer of frontier AI models and a company warning governments about risks that its own technology could create.
Its Responsible Scaling Policy specifically addresses catastrophic risks from increasingly capable AI, while its proposed Advanced AI Framework calls for governments to have authority to block or deter dangerous deployments.
For a country such as Kenya, that raises a question that cannot be answered simply by requiring future projects to comply with Kenyan law: what happens when the risks exceed the ability of Kenyan institutions to manage them?
If a frontier model were to cause a major cybersecurity incident, disrupt critical infrastructure, enable large-scale political manipulation or produce a capability that its developer itself considered potentially catastrophic, Kenya would need more than contractual compliance. It would need independent technical capacity to evaluate the system, access to relevant safety information, emergency powers, cybersecurity expertise and a meaningful role in international AI governance.
Those capabilities cannot be outsourced indefinitely. The asymmetry is particularly critical because Anthropic has argued that AI is advancing faster than governments can regulate it. Its policy framework says AI could eventually accelerate its own development and that increasingly capable systems present risks including cyber operations, biological threats and loss of control.
The United Nations’ independent scientific panel on AI has similarly warned that AI development is moving faster than scientific understanding and government policy, with potential risks including misinformation, cyberattacks, biological threats and loss of control.
That makes local capacity more than an economic-development issue. It is a national-security and sovereignty issue.
Kenya has strong credentials as an African technology hub, particularly in mobile money and startups, but being an early adopter of digital technology is not the same as possessing frontier AI capacity. BCG says Africa risks becoming primarily a consumer of AI rather than a creator unless investment in data centres, computing infrastructure, skills and other digital foundations accelerates.
The Kenyan government faces that difficult policy test as it turns the Anthropic framework into actual projects: whether cooperation produces domestic capabilities that remain after a particular foreign model or company changes course, or whether it deepens dependence on a small number of overseas technology providers.
The agreement’s provisions requiring future data-related projects to comply with Kenyan law are an important starting point. But data residency alone does not amount to technological sovereignty. A country can control where its data is processed while remaining dependent on foreign models, chips, cloud infrastructure, safety evaluations and technical expertise.
Kenya is preparing for a major election in 2027 while simultaneously expanding cooperation with a company whose own threat-intelligence team has documented the use of its model in a Kenyan influence operation.
The partnership could become a test of whether Kenya can move from being an adopter of frontier AI to an institution capable of independently auditing, governing and, when necessary, rejecting it.
Stay ahead of the Stories shaping our world. Subscribe to Impact Newswire and join our
WhatsApp Channel for updates on global tech, business, and innovation—all in one place.
Dive deeper into the future with the Cause Effect 4.0 Podcast, where we explore the ideas, trends, and technologies driving the global AI conversation.
Got a story to share? Contact Us to reach a global audience with Impact Newswire.
Mohd Hassan has extensive experience in news gathering, editing, and writing for the newswire industry, Contact – Info@impactnews-wire.com
Discover more from Impact Newswire
Subscribe to get the latest posts sent to your email.
