named programs · 2 · from sources
What they call their work
AI Safety Research and Policy Publications
Produces research papers and policy guides on topics such as AI alignment, deception, autonomous AI development, and US-China cooperation in AI safety.
International Dialogues on AI Safety (IDAIS)
Flagship event series convening leading scientists from around the world, including from the US and China, to build consensus on AI safety risks and governance interventions.
activities · 1 group
What they do
-
AI Governance and Ethical Research 6 activities
- Advocating for global cooperation on AI safetyAdvances the position that AI safety is a global public good by facilitating dialogue between Western and Chinese scientists and promoting international coordination among key actors to minimize extreme AI risks.
- Conducting research on AI risks and governanceConducts technical and policy research on AI safety, including risks of AI deception, loss of control, autonomous AI development, and international cooperation in technical safety. Produces position papers, primers, and analyses to inform governance and developer practices.
- Organizing International Dialogues on AI Safety (IDAIS)Convenes leading scientists and academics globally through the International Dialogues on AI Safety (IDAIS) event series to build consensus on risks from frontier AI systems and necessary governance interventions. Events have been held in Venice, Beijing, and Oxford, resulting in public consensus statements and policy guidance.
- Producing consensus-based policy guidance and public statementsDevelops and maintains a dynamic policy guide based on consensus from IDAIS meetings, translating expert agreement into actionable recommendations for stakeholders. Produces public statements signed by participants outlining shared goals for AI safety and governance.
- Providing advisory services on AI risk mitigationOffers advisory work on mitigating risks from artificial intelligence to organizations and stakeholders aligned with its mission, including non-profits and international actors.
- Running workshops on AI safety and governanceConducts workshops focused on key topics in AI safety and governance, facilitating expert discussion and knowledge sharing on mitigating risks from advanced AI systems.
-
financials · form 990 · fy2024
revenue
Total revenue$2.18M
Contributions & grants$2.18M100%
Program service revenue$00%
Investment income$00%
Other revenue$0
expenses
Total expenses$52K
Program expenses0%
Admin / overhead100%
Fundraising0%
Salaries & benefits$39K
Grants paid out$0
Largest expense lineCompensation
balance sheet
Total assets$2.13M
Cash$2.13M
Investments$0
Liabilities$3K
Net assets$2.13M
Liquid reserves493.0 mo
1 years on record · 2024–2024
leadership · form 990 part vii · fy2024
Who runs it
paid leadership · 2
| Name | Title | Hours/wk | Compensation |
|---|---|---|---|
| Conor McGurk | Managing Dir. | 1 | $7K |
| Fynn Heide | Executive Dir. | 1 | $7K |
board members · 1
- Adam Gleave — Audit Committee
relationships · 15
Who they work with
- FAR.AI Partner — Former fiscal sponsor of Safe AI Forum prior to its transition to independent 501(c)(3) status.
- Gov.AI Partner — Collaborated on research related to AI governance and safety dialogues between the US and China.
- IDAIS–Shanghai Partner — Provided background context and supplementary explanation for the IDAIS–Shanghai Consensus Statement.
- International Dialogues on AI Safety Partner — Collaborates on organizing the IDAIS events series focused on AI safety.
- International Dialogues on AI Safety Partner — Collaborative initiative convened by the organization to build global consensus on AI safety risks and governance.
- International Dialogues on AI Safety (IDAIS) Partner — Flagship event series that brings together scientists globally to address extreme AI risks.
- International Dialogues on AI Safety (IDAIS) Partner — Flagship series of international dialogues hosted by Safe AI Forum to discuss extreme AI risks and promote global cooperation.
- Oxford Martin AI Governance Initiative Partner — Collaborated on research related to AI governance and safety dialogues between the US and China.
- Yoshua Bengio Partner — Co-author on a research paper about international cooperation in technical AI safety.
- Yoshua Bengio Partner — Collaborated on research regarding cooperation between geopolitical rivals in technical AI safety.
- arXiv Network — Hosts the full text of the organization's research paper on its platform.
- 人工智能安全国际对话 Partner — 为《人工智能安全国际对话上海共识声明》提供研究支持和背景说明。
- 北京 Partner — 举办人工智能安全国际对话（IDAIS）的地点之一，用于达成治理共识
- 威尼斯 Partner — 举办人工智能安全国际对话（IDAIS）的地点之一，用于达成治理共识
- 牛津 Partner — 举办人工智能安全国际对话（IDAIS）的地点之一，用于达成治理共识
strategies · 4
How they approach the work
Named approaches extracted from this org’s sources. Where others share an approach, follow it to see the full set of orgs running it.
- Common-Ground Diplomacy in US-China AI Cooperationmethodology: common-ground-focused diplomacyBy focusing on politically neutral, shared concerns in AI safety and avoiding contentious topics, we increase the likelihood of productive bilateral cooperation because mutual trust and actionable coordination are more sustainable.
- Consensus-to-Policy Translation for Global Stakeholdersmethodology: policy-translationBy building expert consensus on AI risks and translating it into actionable policy guidance, we enable practical implementation across diverse stakeholders because bridging technical understanding and policy action increases adoption of safety measures.
- Evidence-Informed Risk Analysis and Safeguards Developmentmethodology: empirical-risk-analysisBy analyzing empirical evidence on AI deception and shortcomings in alignment, and identifying specific risks in autonomous AI R&D, we guide safer development practices because targeted safeguards are more likely to prevent loss of control in frontier systems.
- Global Collaborative Governance for AI Risk Reductionmethodology: global-collaboration-for-ai-governanceBy fostering shared understanding and collaboration among key global actors—especially in science and policy—we reduce catastrophic AI risks because coordinated governance is more likely to produce aligned, effective safeguards across borders.