概述
They are not binding law, but they shaped the G20 AI Principles, the EU AI Act's definition of an AI system, and many national AI strategies.
深入探討
The OECD Council adopted its Recommendation on Artificial Intelligence in May 2019. Member countries and several non-members adhered to it, and in June 2019 the G20 adopted AI principles drawn from it, extending its reach to major economies outside the OECD. The five values-based principles for trustworthy AI are: inclusive growth, sustainable development and well-being; respect for the rule of law, human rights and democratic values, including fairness and privacy; transparency and explainability; robustness, security and safety; and accountability. Alongside them are five recommendations for policymakers: invest in AI research and development; foster an inclusive AI-enabling ecosystem; enable an interoperable governance and policy environment; build human capacity and prepare for labour market transition; and cooperate internationally. The OECD revised its definition of an AI system in November 2023. It now describes a machine-based system that, for explicit or implicit objectives, infers from its input how to generate outputs such as predictions, content, recommendations or decisions that can influence physical or virtual environments, noting that systems vary in autonomy and adaptiveness. The EU AI Act's definition closely follows this wording, which helps companies and regulators speak a common language. In May 2024 the principles themselves were updated to address general-purpose and generative AI. Changes included stronger attention to misinformation and disinformation, safety, including the ability to override or safely decommission systems that behave in unwanted ways, environmental sustainability, and interoperability between governance regimes. The OECD supports the principles with the OECD.AI Policy Observatory, launched in 2020, which tracks national policies and data, and with work such as incident monitoring. A common misconception is that the principles impose obligations on companies. They are a government-level recommendation; their influence comes through national laws, standards and procurement that adopt them.
戰略影響
風險與安全
災難性和日常的人工智慧危害都取決於誰了解風險以及誰能夠採取行動。
更明確的決策
民眾和專業素養決定強而有力的安全政策在政治上是否可行。
突破炒作
清晰的解釋可以減少炒作、實驗室公關和模糊道德劇場的影響。
The Future of OECD AI Principles
The OECD's role has grown as a coordination point rather than a regulator. It supports reporting under the G7 Hiroshima AI Process and works on incident tracking and measurement, which could give governments more comparable evidence. Because the principles are high level, their practical effect depends on whether national laws, standards bodies and procurement rules keep referring to them. Divergence between regional approaches, such as the EU's binding regulation and lighter-touch regimes elsewhere, makes a shared baseline useful, but it does not guarantee convergence. Further revisions are likely as AI capabilities change.
現實世界的實施
EU lawmakers aligned the AI Act's definition of an AI system with the OECD's updated 2023 definition, so the two frameworks describe the same kinds of systems.
A national government writing its AI strategy organises its commitments around the OECD's five recommendations, such as investing in AI research and preparing workers for labour market changes.
A policy researcher uses the OECD.AI Policy Observatory to compare how different countries fund AI research and regulate AI.
A company's responsible AI policy maps its internal rules to the OECD principles of transparency, robustness and accountability to show customers its approach follows an international baseline.
風險與防護欄
將存在風險視為科幻小說，同時能力複合。
混淆了表面產品安全與高度自治下的對準。
只給非英語和非專業觀眾留下低品質的資源。
實施路線圖
單獨的產品危害、誤用和失控/失調風險。
詢問哪些證據會改變您對時間表和嚴重性的看法。
比起行銷主張，更喜歡主要來源和具體評估。
確定一條行動路徑：職業、政策、資金或技能——而不僅僅是意識。
不斷探索
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the OECD AI Principles quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
常見問題
What is OECD AI Principles?
The OECD AI Principles, adopted in May 2019, were the first intergovernmental standard on artificial intelligence: five values-based principles for trustworthy AI and five recommendations for governments. They are not binding law, but they shaped the G20 AI Principles, the EU AI Act's definition of an AI system, and many national AI strategies.
When were the OECD AI Principles first adopted?
The OECD Council adopted its Recommendation on AI in May 2019, the first intergovernmental AI standard.
Which of these is one of the five values-based principles?
The five are inclusive growth and well-being; rule of law, human rights and democratic values; transparency and explainability; robustness, security and safety; and accountability.
Which item is a recommendation for policymakers rather than a values-based principle?
Building human capacity is one of the five recommendations to governments; the others listed are values-based principles.
Which body adopted AI principles in June 2019 drawn from the OECD's?
The G20 AI Principles drew from the OECD Recommendation, extending its reach beyond OECD members.
What word in the OECD definition most distinguishes AI systems from simple rule-based software?
The definition says an AI system infers from inputs how to generate outputs, which is key to most readings of scope.
繼續學習
相關指南
為此主題精選的更多指南