International AI Safety Report 2026 Launched at Daystar University

Policy leaders, tech experts, and academics gather at Daystar University’s Nairobi Campus during the official launch of the International AI Safety Report 2026 on September 9, 2026.

By Timothy Oiro

[Nairobi, September 9, 2026] The International AI Safety Report 2026, published by the UK Government last February through the AI Security Institute (AISI) was launched at Daystar University’s Nairobi Campus, Kenya the fifteenth country globally to launch the report.

The launch event attracted participation from key decision makers, policy leaders, tech experts, and academia, led by the British High Commissioner to Kenya, H.E. Matt Baugh.

Highlighting the ongoing UK-Kenya partnership in science and technology, H.E. Mr. Baugh called for global cooperation in a bid to address emerging technological risks.  He lauded the UK Government investing heavily in AI through AISI. 

The Kenya Government was represented by the Secretary of ICT at the State Department for ICT and Digital Economy, Mr. Emmanuel Kata Kimeu.   

Mr. Kimeu emphasized that AI safety can no longer be treated as a distant concern, saying the new report, which he termed a timely milestone, grounds policy decisions in concrete scientific evidence while keeping human safety at the centre.

Others present included the CEO of Africa Advanced Level Telecommunications Institute (AFRALTI), Mr. William Baraza,  Daystar University Vice-Chancellor Prof. Laban Ayiro, DVC-Academic, Research and Student Affairs, Prof. Samuel Muriithi, and DVC-Finance, Administration and Planning Prof. Muturi Wachira, Management Board members, faculty, staff and students. The lead writer Dr. Carina Prunkl presented the report virtually from France.

In his remarks, Prof. Laban Ayiro expressed concern that despite achieving global internet connectivity, Africa’s output of scientific research stands at only three per cent. He further emphasized that artificial intelligence is no longer just a simple tool, but an operating environment for daily work, human judgment, security, and social truth.

"Universities must step up the effective use AI in teaching and research, while helping to shape public imagination," said Prof. Ayiro. He called for a strong moral architecture grounded in dignity, stewardship, oversight, and accountability to guide the technology going forward.

Reinforcing this commitment, the Deputy Vice-Chancellor for Academic, Research and Student Affairs Prof. Samuel Muriithi, highlighted the strides Daystar University has made to engage with Artificial Intelligence.  He cited the recent DU-BRAINS AI Conference held in July, saying the university was constantly building the capacity of faculty to facilitate teaching and integration AI into higher education.

The lead report writer, Dr. Carina Prunkl, lauded Kenya’s active participation in global policy formulation that ensures that Global South perspectives shape the future of international AI safety policies.

Speaking about the AI Evidence Dilemma, Dr. Prunkl outlined the following key trends, risks, and defense strategies detailed in the 2026 report:

• Growing Capabilities: AI developers have increased the computing power used to train leading models by roughly five times each year. If current trends hold, AI systems could operate autonomously on multi-day tasks by 2030 with the duration of software engineering tasks that AI agents can complete doubling every seven months.

• Key Risks Identified: The report breaks down general-purpose AI threats into three distinct areas:
  • Misuse: Scams, cyber attacks, user manipulation, and potential misuse in biological or chemical domains.
  • Malfunctions: System failures, unsafe advice, and unreliable computer code.

  • Systemic Impact: Widespread disruptions to labor markets and threats to human autonomy and social trust.

• Safety & Resilience Measures: Managing AI threats requires pre-deployment testing, real-time monitoring, and building societal resilience. Dr. Prunkl further noted the unique challenge of "open-weight" models, saying that while they foster innovation, their safety features can be easily removed, and once released, a model code cannot be recalled.

The day concluded with a panel discussion moderated by the Director of Design-for-Delight (D4D) Mr. Fred Ogore. The panel discussion featured insights by Dr. Stephanie Kasaon, the AI Trust and Safety Lead at Action Lab Africa, and Ms. Gathoni Njeri, a research scholar from ILINA Program that advances AI safety, governance and responsible AI development globally.  The panelists explored how the report's findings apply directly to national policy across Kenya and the wider Global South.

The event truly underscored Daystar’s commitment to supporting informed policy making by identifying the risks in AI adoption particularly general-purpose AI models and systems that are designed to perform a variety of tasks, rather than one specialized function.