Frontier AI Safety Commitments
In plain wordsPromises made in 2024 by major AI companies to publish safety plans for their most advanced models, including the danger points at which they would stop if the risks could not be controlled.
Voluntary commitments announced at the AI Seoul Summit on 21 May 2024 by 16 AI developers from North America, Europe, Asia and the Middle East — among them Amazon, Anthropic, Google, IBM, Meta, Microsoft, Mistral AI, Naver, OpenAI, Samsung Electronics, xAI and Zhipu AI — later joined by Magic, MiniMax, 01.AI and NVIDIA, for 20 in all. Each signatory accepted three kinds of obligation: to manage the risks of its frontier modelsfrontier model, n.: One of the most capable AI models in existence at a given time, built by a handful of leading labs at great cost, and the main target of AI safety rules. throughout their development and use, defining in advance the level of severe risk it would refuse to tolerate without safeguards and, as a last resort, not building or releasing a model whose risks it could not bring below that level; to answer for this through internal governance; and to be open about its approach with governments and the public. Each company promised to publish a safety framework before the Paris AI Action Summit of February 2025, and most did, as a responsible scaling policyresponsible scaling policy, n.: A public rulebook in which an AI company promises to test each new model for dangerous abilities and to add stronger safeguards, or slow down, as its models become more capable. or similar document built on capability thresholdscapability threshold, n.: A level of dangerous ability — such as meaningfully helping someone make a biological weapon — that, once an AI model reaches it, obliges its developer to add stronger safeguards or hold the model back. and if-then commitmentsif-then commitment, n.: A promise by an AI developer of the form: if our model can do a particular dangerous thing, then we will have these protections in place — or we will stop..
Under the Frontier AI Safety Commitments, the lab published a framework naming its intolerable-risk thresholds.
Origin
Brokered by the UK and South Korean governments as hosts of the Seoul summit, building on the voluntary commitments that US developers made to the White House in July 2023 and on early company safety policies such as Anthropic's responsible scaling policy (September 2023).
Usage note
Non-binding and self-defined: each company sets its own thresholds and judges its own compliance, which critics say makes them hard to verify. California's SB 53SB 53, prop. n.: A 2025 California law requiring developers of the most powerful AI models to publish their safety plans, report serious safety incidents to the state and protect employees who raise alarms. later made publishing such frameworks a legal duty for large frontier developers.
Related
- responsible scaling policyresponsible scaling policy, n.: A public rulebook in which an AI company promises to test each new model for dangerous abilities and to add stronger safeguards, or slow down, as its models become more capable.
- Frontier Safety FrameworkFrontier Safety Framework, prop. n.: Google DeepMind's published plan for spotting when its AI models approach dangerous abilities and applying stronger security and safeguards before that happens.
- Preparedness FrameworkPreparedness Framework, prop. n.: OpenAI's published process for testing its AI models for catastrophic risks — such as help with biological weapons or cyberattacks — and deciding what safeguards are needed before release.
- capability thresholdcapability threshold, n.: A level of dangerous ability — such as meaningfully helping someone make a biological weapon — that, once an AI model reaches it, obliges its developer to add stronger safeguards or hold the model back.
- if-then commitmentif-then commitment, n.: A promise by an AI developer of the form: if our model can do a particular dangerous thing, then we will have these protections in place — or we will stop.
- AI Safety SummitAI Safety Summit, prop. n.: A series of international meetings of governments, AI companies and researchers, begun in England in 2023, to agree on how to handle the risks and benefits of the most advanced AI.
- frontier AIfrontier AI, n.: The most capable general-purpose AI systems of the moment — those at or beyond the leading edge — which governments single out for extra safety attention.
Sources
- UK Department for Science, Innovation and Technology (2024). Frontier AI Safety Commitments, AI Seoul Summit 2024. 21 May 2024; signatory list updated 7 February 2025. gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024/frontier-ai-safety-commitments-ai-seoul-summit-2024
Mentioned in
AI governanceAI governance, n.: The rules, institutions and habits that shape how AI is built and used — from laws and treaties to technical standards and a company's own review processes., AI Safety SummitAI Safety Summit, prop. n.: A series of international meetings of governments, AI companies and researchers, begun in England in 2023, to agree on how to handle the risks and benefits of the most advanced AI., frontier AIfrontier AI, n.: The most capable general-purpose AI systems of the moment — those at or beyond the leading edge — which governments single out for extra safety attention., Bletchley DeclarationBletchley Declaration, prop. n.: A short statement signed by 28 countries and the European Union at the first global AI safety summit, held in England in 2023, agreeing that the most advanced AI could cause serious harm., capability thresholdcapability threshold, n.: A level of dangerous ability — such as meaningfully helping someone make a biological weapon — that, once an AI model reaches it, obliges its developer to add stronger safeguards or hold the model back., frontier labfrontier lab, n.: One of the handful of companies building the most advanced AI models of the moment, spending billions of dollars on computing power to train them., G7 Hiroshima AI ProcessG7 Hiroshima AI Process, prop. n.: An effort begun by the G7 group of wealthy democracies in Japan in 2023 that sets voluntary rules for companies building the most advanced AI, which now report publicly on how they follow them., General-Purpose AI Code of PracticeGeneral-Purpose AI Code of Practice, prop. n.: A voluntary rulebook, published by the EU in 2025, that makers of large AI models can sign to show they follow the EU AI Act's rules on transparency, copyright and safety., kill switchkill switch, n.: A way to quickly shut down an AI system, or halt its development, if it starts behaving dangerously — anything from an off button to a company's promise to stop., responsible scaling policyresponsible scaling policy, n.: A public rulebook in which an AI company promises to test each new model for dangerous abilities and to add stronger safeguards, or slow down, as its models become more capable., Frontier Model ForumFrontier Model Forum, prop. n.: An industry group set up in 2023 by leading AI companies to share safety research and best practices for the most advanced AI models., if-then commitmentif-then commitment, n.: A promise by an AI developer of the form: if our model can do a particular dangerous thing, then we will have these protections in place — or we will stop., MolochMoloch, prop. n.: A name, borrowed from an ancient god of sacrifice, for situations where competition forces everyone to act against their shared interests — as when AI companies feel pushed to race ahead despite the risks.