Cambridge AI Safety Hub

What Cambridge AI Safety Hub does

Cambridge AI Safety Hub (CAISH) is a non-profit field-building organization founded in 2024 and based in Cambridge, UK, operating as a fiscally sponsored project of Meridian Impact CIC (a UK Community Interest Company, company number 13653958) and funded by Coefficient Giving. Its mission is to expand the AI safety talent pipeline by bringing promising students and early-career professionals into the field through education, community, and mentored research. CAISH does not build a proprietary technical product; its core outputs are structured programmes including the Alignment Fellowship (five-week introductory course with technical and governance tracks), Alignment Desk (writing accelerator producing three published pieces per participant per term), MARS (part-time research programme pairing 2–4 person teams with mentors from Redwood Research, Google DeepMind, UK AISI, MIRI, and others), the Hardware Assurance Programme (six-day intensive bootcamp), the UK AI Policy Course, and the AI Workload Verification reading course. All programmes are free of charge to participants, with CAISH covering travel, accommodation, meals, compute budgets (~$2k+ per MARS team), and stipends (e.g., £1,500 for the 20-person Hardware Assurance cohort).

CAISH's go-to-market is community-led and geographically concentrated in Cambridge, relying on a student committee, word-of-mouth, events on Luma, and direct outreach within the Cambridge academic and professional ecosystem rather than paid marketing. Distribution is direct: in-person delivery at the Meridian office with hybrid remote phases for MARS. The customer base is segmented into Cambridge-based students (undergrads, PhDs, Masters) seeking introductory context and peers, professionals exploring career transitions into AI safety (including hardware engineers and policy workers), and the broader Cambridge professional AI safety community attending events and socials. Pricing is uniformly free, eliminating price-based segmentation entirely.

The business model is fully philanthropic: there are no commercial revenue streams, and CAISH is funded by grants from Coefficient Giving under fiscal sponsorship by Meridian Impact CIC. The organization is led by co-directors Justin Dollman and Gaurav Yadav (succeeding founding MARS director Chloe Li, now an Anthropic Fellow). The organization operates a portfolio of brand-level programmes (Alignment Fellowship, Alignment Desk, MARS, Hardware Assurance Programme, UK AI Policy Course, AI Workload Verification Course) and maintains deep partnerships with leading AI safety research organizations that supply mentors and research collaboration.

Name

Cambridge AI Safety Hub

Legal name

Cambridge AI Safety Hub (operating as a project of Meridian Impact CIC)

Website

https://caish.org

Company type

Private

Founded year

2024

Operating status

Operating

Headcount range

11–50 employees

Short description

Cambridge-based non-profit field-building organization that educates students and professionals in AI safety through free fellowships, mentorship-based research, writing accelerators, and hardware assurance bootcamps, funded by Coefficient Giving under fiscal sponsorship by Meridian Impact CIC.

Ownership category

akta.pro rank

NAICS

Computer Training (611420)

akta.pro primary industry

Responsible AI, AI Governance & Compliance Services (BPAEAHAJ)

akta.pro secondary industries

Safety, Alignment & Content Moderation (Guardrails, Red Teaming) (HDAAACAI), AI Observability, Monitoring & Evaluation Platforms (Drift, Quality, Safety) (HDAEANAF), Regulatory Readiness & Audit Automation (e.g., EU AI Act, NIST AI RMF, ISO/IEC 42001) (HDAAAMAE)

Keywords

• AI safety education
• AI policy training
• Alignment research mentorship
• Hardware assurance bootcamp
• Existential risk community

Headquarters

HQ city

Cambridge

HQ country

United Kingdom

HQ region

Europe

Offices1 record

Markets served

GTM type

B2C

Offering type

Services

Cost components

Personnel, Operations, Technology or R&D, Marketing or Sales, Infrastructure

Revenue model

1. Grants and philanthropic funding: CAISH is a fiscally sponsored project of Meridian Impact CIC (a UK Community Interest Company, company number 13653958) and is funded by Coefficient Giving. No commercial revenue streams are described. All programmes are offered free of charge to participants. CAISH covers travel, accommodation, meals, and compute costs for participants.

Pricing tiers

Go-to-market motion1 record

Distribution channels1 record

Marketing channels5 records

Core offering

Cambridge AI Safety Hub runs free educational programmes and community-building initiatives to bring students and early-career professionals into AI safety. Its core offerings include the Alignment Fellowship (introductory course), Alignment Desk (writing accelerator), MARS (part-time mentored research programme), UK AI Policy Course, Hardware Assurance Programme (bootcamp), and AI Workload Verification reading course. All programmes are free of charge, with travel, accommodation, meals, compute budgets, and stipends covered for participants.

Product overview

Cambridge AI Safety Hub is a community organization that offers a portfolio of educational programmes focused on reducing existential risks from advanced AI. The core offerings include the Alignment Fellowship (introductory five-week course with technical and governance tracks), UK AI Policy Course (practical three-week policy programme), Alignment Desk (writing accelerator), MARS (mentorship-based research programme), Hardware Assurance Programme (six-day intensive bootcamp), and AI Workload Verification Course (structured reading course). All programmes are free of charge and serve to build the AI safety community in Cambridge by helping students and professionals enter the field.

Differentiator

Problem solved

Functional benefit

Brands

• MARS (Mentorship for Alignment Researchers): Part-time technical AI safety and governance research programme pairing teams with experienced mentors.
• Hardware Assurance Programme
• AI Workload Verification Course
• UK AI Policy Course
• Alignment Desk

Quantifiable outcome

• MARS alumni have published at top venues including NeurIPS 2025, ICLR 2026, ICML 2026, and IEEE S&P 2026.
• +3 more outcomes

Named customers2 records

Segments3 records

Ideal customer profiles3 records

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

Integration6 records

Partnerships

Nine partnerships are on record, tiered major and minor.

• Future of Life InstitutemajorStrategic or Co-development PartnerThe Hardware Assurance Programme features Future of Life Institute as one of the organizations presenting on Day 5, alongside Amodo, Tampersec, Seldon Lab, and Coefficient Giving. Residents include Péter Drótos (Hardware Engineer) and Yannick Mühlhäuser (Policy Researcher) from Future of Life Institute. The Institute works on hardware assurance infrastructure and governance.
• AmodomajorStrategic or Co-development PartnerAmodo Design presents their current work and technical bottlenecks at the Hardware Assurance Programme. Amodo works on Plan A verification (network taps, recomputation capture) and has published research including network traffic hashing and memory wiping. They are a core technical partner in the hardware assurance space.
• TampersecmajorStrategic or Co-development PartnerTampersec presents their work on tamper-evident hardware at the Hardware Assurance Programme Day 5. They are a technical organization building in the hardware assurance space.
• Seldon LabmajorStrategic or Co-development PartnerSeldon Lab presents their hardware assurance work at the Hardware Assurance Programme Day 5. They are among the organizations shaping the direction of the hardware assurance field.
• Redwood ResearchmajorStrategic or Co-development PartnerRedwood Research provides mentors to the MARS programme. Tyler Tracy of Redwood Research mentors a MARS team. Redwood researchers also give talks at the Alignment Fellowship. Redwood Research is a leading technical AI safety organization.
• Google DeepMindmajorStrategic or Co-development PartnerGoogle DeepMind provides mentors to MARS (Rishub Jain and Elizaveta Tennant). Their researchers contribute to the AI safety ecosystem that CAISH draws on for mentorship and talks.
• Institute for AI Policy and Strategy (IAPS)majorStrategic or Co-development PartnerIAPS provides mentors to MARS (Clarissa Koh). Will Hodgkins, a Senior Researcher at IAPS, is cited as the curator of the public reading list for the AI Workload Verification course.
• Centre for the Study of Existential Risk (CSER)minorStrategic or Co-development PartnerCSER is mentioned in the context of Alignment Desk participant Cain Hillier being an MPhil student in Global Risk & Resilience at CSER. CSER is a Cambridge-based research institute focused on existential risk, and its students participate in CAISH programmes.
• BlueDot ImpactminorStrategic or Co-development PartnerBlueDot Impact's Rapid Grants programme (up to $10,000) is referenced as a funding option for Hardware Assurance Programme participants who want to kick off their own projects after the week. Prior exposure through BlueDot is noted as helpful but not required for MARS.

Scale indicators5 records

Recent moves6 records

Expansion highlights5 records

Direct peers

• BlueDot Impact: UK-based AI governance field-builder that runs cohort-based AI safety and biosecurity courses. Directly comparable to CAISH as a non-profit field-building organization delivering introductory AI safety programmes, and explicitly named in CAISH's data as a complementary partner.
• MATS (ML Alignment & Theory Scholars): Mentorship-based research programme for AI alignment, named as a placement destination for CAISH's MARS alumni. Operates a similar mentorship-pairing model for technical alignment research but at larger US-anchored scale.
• AI Safety Camp: Short-form training programmes in AI safety targeting students and early-career researchers. Directly comparable to CAISH's Alignment Fellowship in format and audience, serving as a competing talent pipeline entry-point.
• Global Priorities Institute (Oxford): Oxford-based academic institute focusing on longtermist priorities including AI risk. Comparable as a UK university-anchored AI safety research and training environment serving a similar student demographic to CAISH.
• Centre for the Study of Existential Risk (CSER): Cambridge-based research centre at the University of Cambridge focused on existential risk, including AI risk. Named directly in CAISH's data as a partner whose MPhil students participate in CAISH programmes; serves overlapping Cambridge academic audience.

Regional players

• Conjecture: London-based AI safety research and applied alignment organization. Operates in the same UK AI safety talent ecosystem, sharing mentors and participants with CAISH but primarily serving a London-resident audience rather than Cambridge.

Emerging players

• Future of Life Institute: AI safety and governance organization named explicitly as a CAISH programme partner (Hardware Assurance Programme). Operates overlapping AI safety education and policy work but with broader international scope than CAISH.
• GovAI (Centre for the Governance of AI): AI governance research and training organization named in CAISH's mentor network. Overlaps on AI policy training and career-pipeline work, but with a stronger research orientation and broader international footprint.
• Aporia (AI safety community): AI safety community organization providing meetups, talks, and introductory programming. Comparable in community-building focus and event-driven GTM but operates across multiple cities with a less structured fellowship programme than CAISH.

Broad incumbents

• 80000 Hours: Career guidance organization that advises people on high-impact careers including AI safety. Comparable as part of the broader AI safety field-building ecosystem that channels talent into the field, but operates at much larger scale and broader scope than CAISH.

Market position

Strengths4 records

Weaknesses4 records

Competitive moat4 records

Key risks6 records

Key highlights6 records

Customer concentration

• LinkedInlinkedin.com/company/cambridge-ai-safety-hub

Revenue estimate

Valuation estimate

Number of profiles

Profiles2 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Cambridge AI Safety Hub