You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
Apart Research, together with Poseidon Research and AE Studio, is running a three-day sprint on collusion between AI agents, October 23-25, 2026.
Url: https://apartresearch.com/sprints/ai-collusion-research-sprint-2026-10-23-to-2026-10-25
This research sprint studies collusion between AI agents, in markets and in everyday agent deployments. Alongside engineers and researchers, the sprint is targeting top quantitative finance talent, so participants will likely be based in the world's top financial capitals.
Apart Research has already secured one in-person hub for the hackathon — the Collider in NYC.
In our meeting with Apart, they had expressed a desire to attract “cracked” finance talent from all over the world.
We at AI Safety HK (AISHK) support the vision behind this research sprint, and we have offered to expand its coverage to one of the world's most important financial centers.
Why an in-person hub in Hong Kong?
Hong Kong is an ideal in-person venue for this event.
It is ranked 3rd on the Global Financial Centres Index, with the HK Financial Secretary recently having made an express ambition to take HK to 2nd rank within 15 years.
Hong Kong has overtaken Switzerland as the world’s top cross-border wealth hub. and already serves as the hub for quantitative finance across the Asia-Pacific region.
Many of the world’s top quantitative finance firms have offices in Hong Kong, and tapping into the talent of those firms would make this research sprint a much greater success.
Over the past few months since we commenced operations, AISHK has had significant success in fieldbuilding for AI safety. However, within our network, the vast majority of elite finance talent has had no opening into hands-on AI safety work. Most of these individuals are unlikely to participate in a team-based research sprint remotely, and would appreciate a physical space during the weekend to build and collaborate.
AISHK has offered to assist Apart Research in creating a conducive environment for this talent to have their first practical engagement with AI safety.
We are submitting this grant application to secure accessible, comfortable in-person facilities in Hong Kong, to create a good experience for the city’s top quants.
The minimum funding target would allow us to achieve this basic goal. It would cover advertising and outreach (this is particularly important given the tight timeline) as well as a suitable coworking space, facilities, and other resources needed to make the experience comfortable and productive for participants. This represents the minimum level of support required to host an in-person hub for a research sprint and attract top quants in Hong Kong.
Reaching the target level of funding would also allow us to deepen our collaboration with Apart Research, providing a pathway for promising participants to continue developing their research beyond the sprint.
We have secured in-principle agreement from Apart Research to provide HK’s top-performing team with their own research fellowship. Apart’s research fellowships can typically last 6+ months.
While Apart Research would be supplying the necessary mentorship, compute and other project-relevant infrastructure for the fellowship awardees, AISHK would provide the physical community infrastructure to keep the team motivated and supported during the fellowship. The requested target funding would be a necessary top-up to our pre-existing coworking-space budget which we secured in our previous grant application, and was initially intended to cover space for our core team only.
With the award of an Apart Fellowship for the top-performing local team, this additional financial support would allow us to accommodate the winning team within our workspace and give them a physical base from which to continue their work.
More broadly, it would also be an important first step towards our longer-term ambition of developing a Hong Kong-based hub for AI safety organisations and teams across the wider community.
A. Minimum: $10,000
Our minimum funding goal will allow us to run the in-person Hong Kong hub for Apart Research’s sprint weekend on AI Agent Collusion. Relevant expenditure headings:
i) Coworking space (Sat-Sun): $5000
This represents an upper-end quote (from WeWork) on co-working spaces available in Hong Kong in central neighborhoods that are in close proximity to all the finance offices. We are still shopping around for more affordable yet convenient providers.
ii) Catering: $2500
This covers catering for two days. It will cover two meals a day plus one round of snacks for around 20-25 people.
iii) Advertising: $2000
Apart Research has provided an initial advertising and outreach budget of $2,000 to promote the event to top finance talent.
Given the extremely tight timeline (<6 weeks to the sprint), we may need to scale up quickly if initial traction is slower than expected. This line item is therefore a buffer to run an emergency publicity push.
Concrete measures we can deploy within days include:
Targeted LinkedIn ads to “quantitative researcher”, “trader”, “data scientist”, etc. in Hong Kong. At ~US$5–8 CPC, a US$1,000–1,500 boost could generate ~150–300 qualified clicks to the event page.
Sponsored posts in finance/tech communities (newsletters, Telegram/Discord/Slack groups, alumni networks). A few US$200–500 placements can reach several thousand relevant professionals.
Short online info sessions + referral pushes, promoted via paid social and our existing finance/EA networks.
If Apart’s initial $2,000 ad grant proves sufficient on its own, we will redistribute this amount across the other budget heads (coworking, catering, fellowship support) to improve the overall experience for participants.
iv) Volunteer honorarium: $500
We will task responsible members of our community with prior hackathon organizing experience to act as on-site facilitators for participating teams. This covers a small honorarium across the sprint as appreciation for their contributions.
---------------------
B. Target: $22,000
Our target funding goal extends the minimum ask to also cover seats within our coworking space for the winning Hong Kong team through their Apart Fellowship, so they have a working base in Hong Kong instead of losing momentum once the sprint weekend ends.
We are currently looking for a centrally-located physical space for our core team. We will add this top-up amount to our current budget to get a larger office space so that we can host the winning team alongside our core team and provide them with a proper “AI safety hub” vibe so that they feel motivated to keep working on their research alongside other community members doing similar work.
i) Coworking space seats for the Apart Fellowship team: $12,000
This is based on roughly 8 months of coworking access at $1,500 a month. We expect 8 months to be enough wiggle room for a team to complete the Apart Fellowship, though we are willing to cover the shortfall out of our existing budget in case it isn’t.
AISHK will handle the organisational logistics for Apart Research’s Hong Kong hub, including local outreach, participant coordination, venue arrangements, and on-the-ground support during the sprint.
AISHK is Hong Kong’s first dedicated AI safety fieldbuilding organisation. Since beginning operations, we have focused on creating accessible entry points into AI safety for local talent who may have relevant expertise but no previous experience in the field.
We have developed a growing programme of reading groups, social events, and speaker events designed to help people engage with AI safety in a practical and approachable way. These events have helped us build a network of professionals across Hong Kong’s finance, technology, academic, legal, policy, and startup sectors who are interested in learning more about AI safety.
Our past events are listed here:
We have already built international collaborations with top AI safety organisations. We are supporting the delivery of TARA in Hong Kong and will support the ML4Good bootcamp taking place in October 2026.
We are also helping develop new student societies at elite Hong Kong universities, in order to create a talent pipeline into AI safety for top graduates.
The AISHK core team brings together experience in community building, fintech, banking, policy, and education:
Kenneth Chan is Managing Director of Effective Altruism Hong Kong. Through EAHK, he has organised fellowship programmes, conferences, and other large community-building events.
He is also the founder of GiveWise, an Asia-focused tax-efficient giving platform, and has spent around a decade building high-impact communities. Through his work with EAHK and GiveWise, he has developed a substantial network of trusted contacts in Hong Kong’s finance, trading, and philanthropic communities.
Abeer Sharma has several years of experience building professional networks in Hong Kong’s fintech and Web3 communities. He has participated in fintech and Web3 hackathons and brings experience with professional outreach and participant engagement, particularly among people working in emerging technology and finance.
Chris Barnes is a Risk Technology Product Lead in Hong Kong’s banking sector and has extensive experience and connections across some of Hong Kong’s largest financial institutions.
Aya Lowe is a technology policy and digital safety adviser with fifteen years of experience across the Asia-Pacific region and the Middle East, spanning government, media, and the technology sector. She is also a programme manager for the Singapore AI Safety Hub’s first-ever AI safety fellowship.
We can also draw on volunteers from our wider community who have organised or won hackathons themselves.
The main failure risk for us presently is that we have less time than ideal for outreach and advertising. With fewer than 40 days before the sprint, we need to move quickly to reach top talent in Hong Kong’s quantitative finance and related professional networks and ensure their interest.
If we fail to reach enough of the right participants in time, the Hong Kong hub will have less impact than intended.
All other logistical issues related to organizing the research sprint are fairly straightforward for the team and we are confident in our ability to handle it.
We have received $85,100 for seeding and scaling AISHK, which had previously been operating on a self-funded basis from founders' personal money since December 2025.
Our previous successful grant application can be found here: https://manifund.org/projects/ai-safety-hong-kong-aishk
However, our previous grant application was tightly-scoped to fund all our planned activities for up to a year.
This current collaboration with Apart Research was recently put together on a fairly short notice, and we’re racing against the clock to ensure we deliver it to a high quality. We cannot deliver the required quality presently without an additional grant as our previous grant money has already been committed to our other planned activities and organizational scaling operations.
There are no bids on this project.