Countries struggling to test AI and prepare for its risks will be a focus of the Bridgetown Special Summit on AI Safety on Monday, November 23, 2026. UN Secretary-General António Guterres and Barbados Prime Minister Mia Mottley will co-host the hybrid meeting in Bridgetown, Barbados. The UN announcement does not provide a start time, timezone or official watch link.
About 150 participants are expected. Although the meeting will combine in-person and remote participation, the announcement does not identify a specific venue or explain how the public can follow online. Its confirmed agenda covers international risk management, national AI safety capacity, and access and inclusion.
A framework countries can use at home
The summit will open discussion on Guterres’ proposal for a multilateral AI risk-management framework, backed by independent and credible oversight. Amandeep Singh Gill, the UN special envoy for digital and emerging technologies, described its purpose on Friday as helping countries build their own national AI safety infrastructure—not establishing a treaty or regulation.
Guterres is expected to emphasize that national safety approaches need strengthening, while recognizing that no single model will fit every country, his spokesperson said. His opening remarks will also address AI’s potential benefits, including stronger health systems, scientific discovery, education and hurricane forecasting, alongside its growing risks.
Gill pointed to rising political attention ahead of the meeting. He said 135 leaders raised AI concerns during this year’s UN General Assembly high-level week, compared with three in 2017. A call by Norway and Finland for control of frontier AI models—the most advanced systems—has been signed by 32 states and the European Commission, he added.
Models may be developed in a couple of geographies, “but the risk can travel anywhere and the capacity is not there to deal with those risks.”
Amandeep Singh Gill, UN Special Envoy for Digital and Emerging Technologies
The gap extends beyond building AI
Gill’s warning concerns more than unequal access to the computing power and talent needed to develop AI. Many countries also lack the capacity to test models built elsewhere and prepare for incidents such as cybersecurity breaches, he said. That leaves national preparedness as a separate challenge from whether a country develops models of its own.
To illustrate the imbalance, Gill cited the World Development Report 2026: high-income countries have 500 incident response teams, while low-income countries have six. He used the comparison to underscore how limited the capacity to respond can be in countries exposed to risks from technology developed abroad.
Inclusion on the agenda, military uses excluded
A second roundtable will focus on access, inclusion and linguistic diversity, reflecting developing countries’ concerns about missing out on AI’s benefits. The agenda therefore pairs the ability to manage risks with the ability to benefit from the technology. Military uses of AI will not be discussed.
Expected participants include ministers from Norway and Finland, AI industry leaders, heads of AI safety institutes and members of the UN Independent International Scientific Panel on AI. The United States will be represented by a senior diplomat, and Chinese experts are also expected.
No negotiated outcome is planned. Calls to action and pledges are expected instead, leaving Monday’s meeting focused on discussion and commitments rather than an agreed international rulebook.
Reader comments
Newest comments first. Replies stay oldest first.