The AI Seoul Summit, co-hosted by the United Kingdom and the Republic of Korea on 21 and 22 May 2024, was the second gathering in the international series that began at Bletchley Park in 2023. Its two main outputs were the Seoul Declaration, an intergovernmental statement reaffirming cooperation on AI safety, innovation, and inclusivity, and the Frontier AI Safety Commitments, a set of voluntary pledges made by the companies that build the most capable models.
Sixteen leading AI organizations signed the Frontier AI Safety Commitments at the summit, including Amazon, Anthropic, Google, Meta, Microsoft, OpenAI, and others spanning the United States, Europe, the Middle East, and China. The commitments asked each company to assess the risks its frontier models could pose across their lifecycle, to “set out thresholds at which severe risks posed by a model or system, unless adequately mitigated, would be deemed intolerable,” and to define how those risks would be kept below the thresholds, including the possibility of not developing or deploying a model at all if risks could not be mitigated.
The pledges also covered accountability and transparency. Signatories agreed to establish internal governance with defined roles and resources, and to provide public transparency on how they implement the commitments, while allowing that sensitive details could be shared with governments rather than published where disclosure would itself create risk. A concrete deliverable was attached: companies committed to publish their safety frameworks ahead of the next summit, which was held in France.
As with the Bletchley Declaration before it, these are voluntary commitments, not law. They bind no company in any enforceable sense, and a firm that fails to honor them faces reputational rather than legal consequences. Their importance is that, for the first time, a sizable group of frontier labs collectively put their names to specific, published safety practices, including the idea of pre-defined risk thresholds, which later fed into the safety frameworks the same companies released. The summit thus turned the broad principles of Bletchley into more concrete, if still self-imposed, corporate commitments.