万里长桥 - AIS Collaboration and Field-building to Bridge the East/West Divide
Singapore is wonderful! I just returned from vacation there visiting a friend. What a beautiful city, delicious food, and friendly people! In particular, I met four members of the AI Safety community there. They’ve motivated this post with their insight and enthusiasm. I reached out before my trip on a simple request for feedback on a project, and one connection turned into many, which turned into a strong desire to help this nascent AI Safety community flourish.
We desperately need more international coordination in AI Safety, especially between China and the US. Singapore (SG) is poised to support that mission. The Garden City has a perfect mix of talent, opportunity, location, and timing. There is tremendous opportunity for both local growth and for US/EU/CN orgs to coordinate with the region.
Unique Benefits
The US and China are the current major players in the AI race, and SG is delicately positioned as a bridge nation between these powers. With a long history of thoughtful diplomacy, western and eastern nations are more comfortable meeting in this neutral ground for sensitive topics such as AI. Verification and evaluation efforts based in SG can reach both Chinese and US labs, solving for a critical piece of the international coordination puzzle. Concordia, the premier Chinese AIS organization, has a very positive outlook on SG’s international position.
2027 will see SG step up to the ASEAN Chair position. With most members of the coalition woefully behind on the AI conversation, this is a tremendous opportunity to build AIS technical capacity in the governments of the region. As the middle powers around the world wake up to their role in the AI conversation, it is extremely timely to support them with the knowledge and experience we’ve built up over the past years.
Fertile Soil
Cut out of the equatorial jungle, the Garden City has been a hotbed of growth for technology, finance, and trade. The AI sector is no different, and AI Safety specifically is ready to bloom.
Most of the AI Safety community in the Bay Area have roots in the tech sector, STEM education, or adjacent. It’s much easier to recruit someone with a baseline of technical knowledge and an engineering brain to the AIS field. SG has a similar talent pool, with top tier universities and experienced tech professionals.
There are some existing safety orgs:
SASH - Verification research, fellowship program, fieldbuilding events, co-working space
I met some of the people in these orgs - they are energized, excited, and ready to make positive impacts. There is also growing public sentiment and government focus toward Responsible AI. What the ecosystem needs is fieldbuilding experience, expert collaboration, and funding. With these inputs, the SG AI Safety community can grow rapidly and make outsized impacts regionally, then globally.
Bridging the Divide
Direct engagement between the West and China runs into many hurdles. Singapore provides a low-friction, low-risk environment to bypass or mitigate some of these hurdles:
Easy immigration/visa requirements from both sides
Socially/politically acceptable engagement
English/Mandarin bilingual society
Mix of Eastern pragmatism and Western EA style thinking
The current AIS work in SG is focused on Verification and Evals. This is great because we need internationally operable kits, acceptable to East and West, in these areas. Metr is great but we also need NEO!
Plus, it’s just a great city to hang out in. Having an attractive physical environment has very real impacts on people’s willingness to work there, whether it’s a long term move or weekend workshop. I can’t say enough good things about the hawker stalls!
Specialization
Having pleaded the case for SG as a great fit for AIS growth, it’s prudent to address a core difference from SF and London. The general mentality and priorities around AI Safety in SG are much more short term and pragmatic. The strongest driver is how to adopt AI capabilities for commercial applications while ensuring business compliance. Responsible AI, another focus area, is along a similar vein, with a higher focus on public protection in areas like privacy/discrimination/transparency. There is relatively little focus on x-risk, biotech, power concentration, or disempowerment. The reasonable approach here seems to be specialization. Not every hub in the world needs to work on all the risk areas. If SG makes contributions in international verification, evals, and public protection, that’s a huge win!
Next Steps
The risk profile is urgent, and Singapore’s position is very timely. Here are a few specific actionables:
Collaborate on projects that require international access/perspective (verification, evals)
Residency program - bring AIS experts to SG for 3-6 months
Have online social meetups to get connected
Field-building
Localize some of the commstools that have been effective in the West, and deploy in SG
Found/fund courses/workshops/fellowships - on-ramps for AIS
University clubs/events at the top SG schools
Collaborate on policy recommendations
Get multi-lateral input on workable international policy for China/US
Engage with SG government through local orgs to build up their technical capacity and provide policy advising
If we believe that international cooperation is essential to long term AI Safety, then Singapore is one of the most important places to invest our energy toward. Let’s build the 万里长桥 10,000 mile bridge together!
万里长桥 - AIS Collaboration and Field-building to Bridge the East/West Divide
Singapore is wonderful! I just returned from vacation there visiting a friend. What a beautiful city, delicious food, and friendly people! In particular, I met four members of the AI Safety community there. They’ve motivated this post with their insight and enthusiasm. I reached out before my trip on a simple request for feedback on a project, and one connection turned into many, which turned into a strong desire to help this nascent AI Safety community flourish.
We desperately need more international coordination in AI Safety, especially between China and the US. Singapore (SG) is poised to support that mission. The Garden City has a perfect mix of talent, opportunity, location, and timing. There is tremendous opportunity for both local growth and for US/EU/CN orgs to coordinate with the region.
Unique Benefits
The US and China are the current major players in the AI race, and SG is delicately positioned as a bridge nation between these powers. With a long history of thoughtful diplomacy, western and eastern nations are more comfortable meeting in this neutral ground for sensitive topics such as AI. Verification and evaluation efforts based in SG can reach both Chinese and US labs, solving for a critical piece of the international coordination puzzle. Concordia, the premier Chinese AIS organization, has a very positive outlook on SG’s international position.
2027 will see SG step up to the ASEAN Chair position. With most members of the coalition woefully behind on the AI conversation, this is a tremendous opportunity to build AIS technical capacity in the governments of the region. As the middle powers around the world wake up to their role in the AI conversation, it is extremely timely to support them with the knowledge and experience we’ve built up over the past years.
Fertile Soil
Cut out of the equatorial jungle, the Garden City has been a hotbed of growth for technology, finance, and trade. The AI sector is no different, and AI Safety specifically is ready to bloom.
Most of the AI Safety community in the Bay Area have roots in the tech sector, STEM education, or adjacent. It’s much easier to recruit someone with a baseline of technical knowledge and an engineering brain to the AIS field. SG has a similar talent pool, with top tier universities and experienced tech professionals.
There are some existing safety orgs:
SASH - Verification research, fellowship program, fieldbuilding events, co-working space
NEO - Technical research - evals
Singapore Internet Governance Forum - Government body for AIS engagement
OpenNexus - Public policy advocacy for AIS
Verify Foundation - International verification coordination
Concordia (SG branch) - Beijing based AIS nonprofit doing research, comms, policy advocacy
I met some of the people in these orgs - they are energized, excited, and ready to make positive impacts. There is also growing public sentiment and government focus toward Responsible AI. What the ecosystem needs is fieldbuilding experience, expert collaboration, and funding. With these inputs, the SG AI Safety community can grow rapidly and make outsized impacts regionally, then globally.
Bridging the Divide
Direct engagement between the West and China runs into many hurdles. Singapore provides a low-friction, low-risk environment to bypass or mitigate some of these hurdles:
The current AIS work in SG is focused on Verification and Evals. This is great because we need internationally operable kits, acceptable to East and West, in these areas. Metr is great but we also need NEO!
Plus, it’s just a great city to hang out in. Having an attractive physical environment has very real impacts on people’s willingness to work there, whether it’s a long term move or weekend workshop. I can’t say enough good things about the hawker stalls!
Specialization
Having pleaded the case for SG as a great fit for AIS growth, it’s prudent to address a core difference from SF and London. The general mentality and priorities around AI Safety in SG are much more short term and pragmatic. The strongest driver is how to adopt AI capabilities for commercial applications while ensuring business compliance. Responsible AI, another focus area, is along a similar vein, with a higher focus on public protection in areas like privacy/discrimination/transparency. There is relatively little focus on x-risk, biotech, power concentration, or disempowerment. The reasonable approach here seems to be specialization. Not every hub in the world needs to work on all the risk areas. If SG makes contributions in international verification, evals, and public protection, that’s a huge win!
Next Steps
The risk profile is urgent, and Singapore’s position is very timely. Here are a few specific actionables:
If we believe that international cooperation is essential to long term AI Safety, then Singapore is one of the most important places to invest our energy toward. Let’s build the 万里长桥 10,000 mile bridge together!