Two major organizations released distinct frameworks for governing artificial intelligence, each addressing different priorities in how the world should manage AI's rapid advancement. UNESCO is pushing for inclusive, global cooperation through its Readiness Assessment Methodology (RAM) program, which has already engaged over 75 countries in developing national AI policies. Meanwhile, Anthropic, an AI safety company, published a framework calling for governments to have legal authority to block or deter dangerous AI deployments, focusing specifically on catastrophic risks from the most powerful systems.

What's Driving the Push for New AI Governance Models?

The urgency behind these complementary proposals stems from AI's exponential growth. Anthropic noted that just a few years ago, AI models could barely write code. Today, Claude Mythos Preview discovered thousands of high-severity vulnerabilities, including in every major operating system and browser. This acceleration means the stakes for governance have risen dramatically. Anthropic warned that without stronger regulatory frameworks, catastrophic risks could emerge across four domains: biological weapons development, cyber attacks on critical infrastructure, loss of control over AI systems, and automated AI research that could amplify other risks.

UNESCO's approach reflects a complementary priority. Rather than focusing narrowly on catastrophic risk prevention, UNESCO emphasizes ensuring AI benefits all nations equally, not just technologically advanced ones. The organization has positioned itself as a bridge between developed and developing countries, helping governments assess their readiness for AI transformation and identify policy gaps.

How Do These Two Frameworks Address Different Governance Challenges?

The frameworks diverge significantly in their mechanisms and scope. Anthropic's Advanced AI Framework targets what it calls "frontier" AI models, defined as those trained using more than 10²⁵ floating-point operations (FLOPs), developed by companies earning more than $500 million in AI-related revenue or spending more than $1 billion on AI research and development. For these high-capability systems, Anthropic proposes that governments should have authority to block deployments that pose catastrophic risks, with civil penalties tied to global annual revenue that escalate with repeated violations.

UNESCO's RAM program takes a broader, capacity-building approach. As of October 2025, over 75 countries from five continents have initiated or completed RAM assessments. The program helps policymakers identify policy gaps and develop national strategies for AI development and deployment. UNESCO also operates the UN Global Dialogue on AI Governance, mandated by the General Assembly, which aims to ensure that AI governance reflects the priorities of all nations, not just the most technologically advanced.

What Specific Safeguards Does Anthropic Propose?

• Transparency Requirements: Frontier AI developers must test their models, publish summaries of results, and share safety frameworks explaining how they evaluate catastrophic risks. They must also publish regular risk reports describing their overall risk posture and engage independent evaluators.
• Independent Evaluation: At least one qualified independent evaluator must review each frontier developer's evaluations and risk reports. Governments and industry should develop standards for these evaluators and ensure they have adequate funding and access to frontier models.
• Security Standards: Developers must secure their entire development environment, protect against threats from outside and inside the company, describe their security program publicly, and test defenses regularly.
• Enforcement Authority: Governments should have legal power to block or deter deployment of models posing significant catastrophic risk, with safeguards to prevent regulatory overreach.

Anthropic emphasized that these rules should apply only to the most capable systems. The company also noted that it supports existing state laws requiring companies to describe their safety practices publicly, but argues that transparency alone is no longer sufficient given the rapid pace of AI advancement.

How Does UNESCO's Global Approach Differ?

UNESCO's strategy emphasizes building capacity across all nations and addressing equity concerns. The organization works on multiple fronts, including promoting AI for gender equality, ensuring cultural diversity is protected as AI transforms content creation and access, fostering media literacy to address how algorithms reshape information access, and developing ethical frameworks for emerging technologies like neurotechnology.

In 2021, UNESCO adopted the Recommendation on the Ethics of Artificial Intelligence, which guides its work and helps member states develop ethically aligned AI policies. In 2025, UNESCO added the Recommendation on the Ethics of Neurotechnology, setting global principles for how neurotechnology should be developed and used while respecting human rights and dignity.

UNESCO also operates the G20 Technology Policy Assistance Facility, developed in cooperation with the South African G20 Presidency. This platform helps policymakers and stakeholders research, design, and implement ethical, inclusive, and future-ready technology and AI policies.

What Are the Key Differences in Philosophy?

The two frameworks reflect different assumptions about AI governance priorities. Anthropic's approach assumes that frontier AI systems pose unique, catastrophic risks that require government intervention beyond what markets or transparency alone can provide. The company explicitly calls for regulatory "teeth," meaning enforcement mechanisms with real consequences.

UNESCO's philosophy centers on ensuring that AI governance is inclusive and reflects the needs of all nations, particularly developing countries that might otherwise be excluded from setting global standards. UNESCO noted that international cooperation on AI governance is essential to ensure that AI development benefits everyone, not just wealthy nations.

Anthropic acknowledged that its framework is written primarily with the US federal government in mind, but it does not believe Congress should preempt state law unless it enacts a federal law at least as strong as Anthropic's proposal. The company suggested that preemption should be surgical, allowing states to regulate AI on issues like child safety and consumer protection that fall outside specific safety functions.

How to Align with Emerging AI Governance Standards

As these frameworks shape policy conversations globally, organizations and governments can take practical steps to prepare for evolving governance expectations:

• Assess Current Readiness: Use UNESCO's RAM program or similar tools to evaluate your organization's or nation's current AI policies, identify gaps, and develop strategies for responsible AI deployment aligned with international standards.
• Implement Transparency Practices: Document and publish safety evaluations, risk assessments, and security practices for AI systems, even if not yet legally required, to demonstrate commitment to responsible development and align with both frameworks' emphasis on openness.
• Engage Independent Review: Seek external evaluation of AI safety practices and risk management approaches to build credibility and identify blind spots in internal assessments, a requirement under Anthropic's proposal.
• Participate in Global Dialogue: Engage with international forums like the UN Global Dialogue on AI Governance to ensure your nation's or organization's priorities are represented in emerging governance standards and contribute to inclusive policymaking.

What Happens Next in AI Governance?

Both organizations are positioning their frameworks as starting points for broader policy conversations. Anthropic urged policymakers to actively engage on these issues now, noting that AI capabilities will improve rapidly over the coming months and governance needs to keep pace. The company acknowledged that its recommendations will likely spark vigorous debate.

UNESCO continues to expand its RAM program, with policymakers worldwide drawing on its assessments and tools to develop national AI strategies. The UN Global Dialogue on AI Governance, coordinated by UNESCO and the International Telecommunication Union (ITU), remains the primary forum for global conversation on AI governance.

Together, these frameworks represent complementary approaches to a shared challenge: how to govern AI responsibly as it advances rapidly. Anthropic focuses on preventing catastrophic risks from the most powerful systems, while UNESCO ensures that governance processes include all nations and address equity concerns. The coming months will reveal how policymakers integrate these distinct priorities into actual legislation and international agreements.