← Back to Dashboard

AI Safety, Alignment, EA, Governance & Responsible AI Hackathons — 2026

Research Report

Research date: 2026-03-18

1. AI Control Hackathon 2026

Name: AI Control Hackathon 2026
Organiser: Apart Research, co-organized with Redwood Research and UK AISI
Domain/Category: AI Safety / AI Control
Prize Pool: $2,000 cash prize pool across all tracks + fully funded trip to ControlConf Berkeley 2026 for 1st place + Apart Research Fellowship invitation
Dates: March 20-22, 2026
Status: ACTIVE (happening this week)
Format: Hybrid (online + multiple in-person hubs)
Location: Global hubs: San Francisco (Mox), Cape Town, Groningen, Warwick University, Sydney (Mar 21-23), Montreal
Platform: Apart Research sprints platform
URL: https://apartresearch.com/sprints/ai-control-hackathon-2026-03-20-to-2026-03-22
Key Themes/Tracks: Control protocol design, ControlArena challenges, Red teaming & vulnerability research
Notable Sponsors: Redwood Research, UK AISI
Team Size: Not specified (individual or team)
Open to: Anyone — no prior AI safety experience required; researchers, engineers, security professionals, AI safety enthusiasts
Quality Signal: HIGH — Apart Research has run 55+ sprints with 6,000+ participants; Redwood Research is the founding group in AI control; UK AISI is a government body. Second edition of this hackathon.
Notes: Top projects reviewed by expert judges. Outstanding work may be invited for further development, publication, or presentation. Feeds into ControlConf Berkeley (April 18-19).

2. AI Manipulation Hackathon

Name: AI Manipulation Hackathon
Organiser: Apart Research
Domain/Category: AI Safety / AI Manipulation Detection & Defense
Prize Pool: $2,000 in cash prizes
Dates: January 9-11, 2026
Status: Recently ended
Format: Hybrid (online + in-person hubs)
Location: Multiple global hubs
Platform: Apart Research sprints platform
URL: https://apartresearch.com/sprints/ai-manipulation-hackathon-2026-01-09-to-2026-01-11
Key Themes/Tracks: Measuring, detecting, and defending against AI manipulation
Notable Sponsors: Apart Research
Team Size: Not specified
Open to: Anyone (500+ participants, 70+ projects submitted)
Quality Signal: HIGH — Apart Research is a well-established AI safety org; winners guaranteed acceptance to present at IASEAI workshop in Paris (Feb 26, 2026)
Notes: Top projects invited to continue in Apart Fellowship and shared with relevant safety researchers.

3. The Technical AI Governance Challenge

Name: The Technical AI Governance Challenge
Organiser: Apart Research
Domain/Category: AI Governance / Technical AI Safety Infrastructure
Prize Pool: ~$2,000 (typical Apart sprint prize, exact amount not confirmed)
Dates: January 30 - February 1, 2026
Status: Recently ended
Format: Hybrid (online + in-person hubs: Berkeley, Montreal, South Africa)
Location: In-person hubs near Downtown Berkeley BART, Montreal, South Africa
Platform: Apart Research sprints platform
URL: https://apartresearch.com/sprints/the-technical-ai-governance-challenge-2026-01-30-to-2026-02-01
Key Themes/Tracks: Verification protocols, monitoring tools, privacy-preserving compliance proofs, coordination systems for enforceable international AI safety cooperation
Notable Sponsors: Apart Research
Team Size: Not specified
Open to: Engineers, researchers, developers (500+ participants expected)
Quality Signal: HIGH — Addresses real policy infrastructure needs
Notes: Focused on technical tools that enable international AI safety cooperation.

4. HackAPrompt 2.0

Name: HackAPrompt 2.0
Organiser: Learn Prompting
Domain/Category: AI Red Teaming / AI Safety / Prompt Security
Prize Pool: $100,000 total
Dates: Multi-month competition (ongoing in 2025-2026)
Status: Active/Ongoing
Format: Online
Location: N/A (virtual)
Platform: hackaprompt.com
URL: https://www.hackaprompt.com/
Key Themes/Tracks:
PointCrow Track (Beginner) — scenario-based challenges
Tutorial Track (Beginner) — prompt injection and jailbreaking fundamentals
CBRNE Track (Intermediate, $50,000 pool) — eliciting CBRNE information
MATS x Trails Track (Advanced) — indirect prompt injection, agent deception
Pliny Track (open-source dataset)
Notable Sponsors: OpenAI, CATO Networks, Pangea, Telus, Towards AI
Team Size: Individual
Open to: Anyone, 150+ countries, 30K+ competitors, no background required
Quality Signal: HIGH — Second edition; first edition produced published research (NeurIPS paper); OpenAI is a primary sponsor
Notes: World's largest AI red-teaming competition. Multiple prize categories including shortest jailbreak ($40K pool at $500 per card), funniest/strangest jailbreak ($30K).

5. OpenAI Red-Teaming Challenge (gpt-oss)

Name: OpenAI gpt-oss Red-Teaming Challenge
Organiser: OpenAI
Domain/Category: AI Safety / Red Teaming / Vulnerability Discovery
Prize Pool: $500,000 ($50K x 10 winning teams)
Dates: August 5-26, 2025 (concluded)
Status: Recently ended (context for future rounds)
Format: Online (Kaggle)
Location: N/A
Platform: Kaggle
URL: https://www.kaggle.com/competitions/openai-gpt-oss-20b-red-teaming
Key Themes/Tracks: Finding novel vulnerabilities in open-weight LLMs — deceptive alignment, reward-hacking, lying
Notable Sponsors: OpenAI
Team Size: Teams allowed
Open to: Anyone (600+ research teams participated)
Quality Signal: VERY HIGH — OpenAI is a frontier AI lab; $500K prize pool is massive
Notes: Focused on gpt-oss-20b and gpt-oss-120b open-weight models. Future rounds likely given success. Watch for announcements of new editions.

6. SafeBench — AI Safety Benchmarking Competition

Name: SafeBench
Organiser: Center for AI Safety (CAIS), supported by Schmidt Sciences
Domain/Category: AI Safety / Benchmarking
Prize Pool: $250,000 (3x $50K first prizes + 5x $20K prizes)
Dates: Launched March 25, 2024; Submission deadline February 25, 2025; Winners announced April 25, 2025
Status: Recently ended (winners announced)
Format: Online (submit benchmarks)
Location: N/A
Platform: mlsafety.org/safebench
URL: https://www.mlsafety.org/safebench
Key Themes/Tracks: Robustness, Monitoring, Alignment, Safety Applications — create novel AI safety benchmarks
Notable Sponsors: Schmidt Sciences
Team Size: Not specified
Open to: Anyone
Quality Signal: VERY HIGH — CAIS is a leading AI safety org; Schmidt Sciences backing; ~100 submissions received; winning benchmarks (Cybench, AgentDojo) adopted by US AISI, UK AISI, and Anthropic
Notes: Prize is specifically for safety (not capability) benchmarks. Winning benchmarks have real-world impact. Watch for future rounds.

7. IEEE SaTML 2026 Competitions

Name: IEEE SaTML 2026 Competition Track
Organiser: IEEE Conference on Secure and Trustworthy Machine Learning
Domain/Category: AI Security / Trustworthy ML / Adversarial ML
Prize Pool: Not specified (academic prestige + potential awards)
Dates: March 23-25, 2026 (conference dates)
Status: Active (conference this week)
Format: In-person conference with online competition components
Location: Technical University of Munich, Germany
Platform: satml.org
URL: https://satml.org/competitions/
Key Themes/Tracks:
Anti-BAD: Defending against backdoors in LLMs (generation, classification, multilingual tracks)
AgentCTF: Evaluating agentic systems through adversarial/defensive challenges (Berkeley RDI Center)
PET-ARENA: Privacy-preserving database systems CTF (white-box and black-box privacy attacks)
Detecting Manipulations of AI Models in Space Operations (funded by European Space Agency)
Notable Sponsors: IEEE, European Space Agency
Team Size: Varies by competition
Open to: Researchers and practitioners
Quality Signal: VERY HIGH — IEEE is a premier academic venue; SaTML is the top venue for ML security/safety research
Notes: Best paper awards also given. Competitions test real-world adversarial scenarios.

8. Amazon Nova AI Challenge 2026: Trusted Software Agents

Name: Amazon Nova AI Challenge: Trusted Software Agents
Organiser: Amazon
Domain/Category: AI Safety / Trusted AI / Agentic Systems Security
Prize Pool: $250K per selected team (sponsorship) + $250K for top developer team + $100K for 2nd developer team + $250K for top red team + $100K for 2nd red team = ~$3.1M+ total
Dates: Applications opened November 2025; tournaments early 2026 through summer 2026; finals September 2026
Status: Active/Ongoing
Format: Hybrid (tournaments + finals event)
Location: Various (finals location TBD)
Platform: Amazon Science
URL: https://www.amazon.science/nova-ai-challenge
Key Themes/Tracks: Building reliable agentic coding systems (developer teams) vs. probing for failures and security weaknesses (red teams)
Notable Sponsors: Amazon
Team Size: University teams (10 selected)
Open to: Full-time students (undergraduate or graduate) only
Quality Signal: VERY HIGH — Amazon is a major tech company; massive prize pool; Year 2 of program
Notes: $250K per team is sponsorship for research, not just prize money. Focuses on trusted AI agents for software development. Build phase + validation/red-team phase structure.

9. Mila National Hackathon — Safer Conversational AI for Youth Mental Health

Name: Championing AI for Good: Building Safer AI for Youth Mental Health
Organiser: Mila - Quebec Artificial Intelligence Institute
Domain/Category: AI Safety / Responsible AI / Mental Health
Prize Pool: CAD 10,000 in monetary prizes + internship with Mila AI Safety Studio
Dates: March 16-23, 2026 (registration deadline March 8)
Status: Active (happening now)
Format: Hybrid (in-person in Montreal + online)
Location: Montreal, Canada
Platform: Mila
URL: https://mila.quebec/en/event/hackathon-championing-ai-for-good-building-safer-ai-for-youth-mental-health
Key Themes/Tracks: Designing, testing, and documenting technical safeguards for conversational AI in youth mental health contexts
Notable Sponsors: Bell Canada, Buzz HPC, Kids Help Phone (KHP)
Team Size: Not specified
Open to: National (Canada)
Quality Signal: HIGH — Mila is a world-leading AI research institute; Bell Canada and Kids Help Phone are major Canadian organizations
Notes: First hackathon of its kind from Mila focused specifically on conversational AI safety. Public opening conference on March 16.

10. CIVICA EU AI Hack

Name: EU AI Hack
Organiser: CIVICA (Hertie School, IE University, SNSPA)
Domain/Category: AI Governance / AI Policy / EU AI Act
Prize Pool: Not specified
Dates: April 17-19, 2026 (hackathon); Pre-hack evening March 23, 2026
Status: Upcoming
Format: In-person
Location: Berlin, Germany
Platform: CIVICA
URL: https://www.civica.eu/news-events/events/detail/ai-policy-x-innovation-europes-hard-choices-on-ai-implementation
Key Themes/Tracks: Algorithmic bias, data governance, transparent AI systems, EU AI Act implementation
Notable Sponsors: CIVICA European University Alliance
Team Size: Student teams
Open to: Students (primarily CIVICA network — 2,500+ students)
Quality Signal: MEDIUM-HIGH — CIVICA is a consortium of top European universities; student-led but strong institutional backing
Notes: Six-month multi-campus initiative; produces digital prototypes and policy briefs. No cash prizes confirmed but strong networking value.

11. iSAFE Hackathon 2026

Name: Inclusive & Secure AI For Everyone (iSAFE) Hackathon 2026
Organiser: CyberPeace Foundation, with ITU (International Telecommunication Union)
Domain/Category: AI Safety / Cybersecurity / Digital Trust
Prize Pool: Cash prizes (amount unspecified) + cloud credits + incubation support + ITU internships + software licenses
Dates: Registration Nov 20, 2025 - Jan 15, 2026; Finale Feb 8-9, 2026; Winners announced at WSIS Forum July 6-10, 2026 in Geneva
Status: Recently ended (finale completed, winners to be announced at WSIS)
Format: Hybrid (global online build + finale in New Delhi + showcase at Bharat Mandapam)
Location: New Delhi, India (finale); Geneva, Switzerland (winner announcement)
Platform: Unstop + CyberPeace
URL: https://etradeforall.org/events/inclusive-secure-ai-everyone-isafe-hackathon-2026
Key Themes/Tracks: Online safety, digital trust, ethical deployment of generative AI
Notable Sponsors: Google.org, ITU, CyberPeace Foundation
Team Size: Up to 3 members
Open to: Individuals and teams worldwide
Quality Signal: MEDIUM-HIGH — ITU is a UN agency; Google.org backing; CyberPeace is an established NGO
Notes: Winners get startup incubation opportunities. International scope.

12. Society-Centered AI Hackathon 2026 (Duke University)

Name: Society-Centered AI Hackathon 2026
Organiser: Duke University (SCAI initiative)
Domain/Category: Responsible AI / AI Ethics / Society-Centered AI
Prize Pool: Not specified (winning team presents on main conference stage)
Dates: February 12-14, 2026 (part of SCAI Conference)
Status: Recently ended
Format: In-person
Location: Duke University, Fuqua School of Business, Durham NC
Platform: Duke SCAI
URL: https://scai.duke.edu/events/society-centered-ai-hackathon-2026
Key Themes/Tracks: Society-centered AI across various topics
Notable Sponsors: Walu AI, Google Cloud, Patrick J. McGovern Foundation
Team Size: 19 interdisciplinary teams (80+ participants)
Open to: Duke community and broader public
Quality Signal: HIGH — Duke is a top university; strong corporate judges/sponsors; second annual edition
Notes: Judged by industry leaders. Focus on interdisciplinary approaches.

13. AI Safety Fieldbuilding Hackathon

Name: AI Safety Fieldbuilding Hackathon
Organiser: Not specified (AI safety community)
Domain/Category: AI Safety / Field Infrastructure
Prize Pool: Not specified
Dates: March 28, 2026
Status: Upcoming
Format: In-person
Location: London, UK
Platform: Not specified
URL: Referenced in AI Safety Events & Training newsletter
Key Themes/Tracks: Building AI tools to improve operational infrastructure for AI safety field-building (courses, funding programs, local groups)
Notable Sponsors: Not specified
Team Size: Not specified
Open to: Anyone — no prior field-building or AI coding experience required
Quality Signal: MEDIUM — Community organized; novel focus on meta-level infrastructure
Notes: Unique focus on building tools for the AI safety ecosystem itself.

14. Re-Align Challenge at ICLR 2026

Name: Re^4-Align Representational Alignment Challenge
Organiser: Re-Align Workshop organizers
Domain/Category: AI Alignment / Representational Alignment / Neuroscience-ML Bridge
Prize Pool: Not specified (academic recognition; poster/talk at ICLR workshop)
Dates: ICLR 2026 workshops: April 26-27, 2026
Status: Upcoming
Format: Hybrid (leaderboard + workshop presentation)
Location: ICLR 2026 venue (Singapore)
Platform: OpenReview + custom leaderboard
URL: https://representational-alignment.github.io/2026/
Key Themes/Tracks: Blue Teams (show model universality via high representational alignment) vs. Red Teams (highlight model variability); representational comparison measures
Notable Sponsors: ICLR
Team Size: Not specified
Open to: Researchers in ML, neuroscience, cognitive science
Quality Signal: HIGH — ICLR is a top-tier ML venue; third edition of the workshop
Notes: Challenge reports submitted via OpenReview. Accepted reports presented as posters/talks.

15. IEEE IES Generative AI Challenge 2026

Name: IEEE IES Generative AI Challenge 2026
Organiser: IEEE Industrial Electronics Society
Domain/Category: Responsible AI / AI Ethics in Industry
Prize Pool: Not specified (opportunity to present at IRAI 2026 or IECON 2026)
Dates: February - April 2026 (2.5 months)
Status: Active
Format: Online
Location: N/A
Platform: IEEE
URL: https://ai.ieee-ies.org/
Key Themes/Tracks: Generative AI in industrial settings: algorithms, models, applications, ethics, regulation and adoption
Notable Sponsors: IEEE
Team Size: Not specified
Open to: Anyone
Quality Signal: HIGH — IEEE is a premier professional organization
Notes: Winners get published paper presentation opportunity at major IEEE conferences.

16. ISB-Punjab Police-CyberPeace Cybersecurity & AI Safety Hackathon 2026

Name: India's Premier Cybersecurity & AI Safety Hackathon 2026
Organiser: Indian School of Business (ISB), Punjab Police, CyberPeace
Domain/Category: AI Safety / Cybersecurity
Prize Pool: Not specified
Dates: 2026 (specific dates TBD)
Status: Upcoming/Announced
Format: Likely hybrid
Location: India
Platform: Not specified
URL: https://www.streettimes.in/isb-punjab-police-cyberpeace-launch-indias-premier-cybersecurity-ai-safety-hackathon-2026/
Key Themes/Tracks: Digital security, AI Safety, emerging cyber threats
Notable Sponsors: ISB, Punjab Police, CyberPeace
Team Size: Not specified
Open to: Students, researchers, technologists, law enforcement, policy experts across India
Quality Signal: MEDIUM — ISB is a top Indian business school; government partnership
Notes: National initiative; cross-disciplinary focus.

17. CISPA European Cybersecurity & AI Hackathon Championship

Name: CISPA European Cybersecurity & AI Hackathon Championship
Organiser: CISPA Helmholtz Center for Information Security
Domain/Category: AI Security / Cybersecurity
Prize Pool: Not specified
Dates: Multiple regional rounds across Europe in 2026; Grand Finale in St. Ingbert, Germany
Status: Active/Ongoing
Format: In-person (regional rounds + finale)
Location: Multiple European cities; Grand Finale in St. Ingbert, Germany
Platform: CISPA
URL: https://cispa.de/en/cispa-hackathon-championship
Key Themes/Tracks: Cybersecurity and AI security challenges
Notable Sponsors: CISPA (Helmholtz Association)
Team Size: Teams
Open to: European participants
Quality Signal: HIGH — CISPA is a leading cybersecurity research center; Helmholtz backing
Notes: Multi-stage championship format across Europe.

18. USAII Global AI Hackathon 2026

Name: USAII's Global AI Hackathon 2026
Organiser: United States Artificial Intelligence Institute (USAII)
Domain/Category: Responsible AI / AI Ethics / AI for Social Good
Prize Pool: $15,000 total ($6,000 high school category, $9,000 college category)
Dates: June 2026
Status: Upcoming
Format: Online (fully virtual)
Location: N/A
Platform: aihackathon.usaii.org
URL: https://aihackathon.usaii.org/
Key Themes/Tracks: Education, productivity, sustainability, health, social impact — with responsible AI principles
Notable Sponsors: USAII
Team Size: Collaborative teams
Open to: High school (grades 9-12), undergraduate, and graduate students worldwide
Quality Signal: MEDIUM — USAII is a professional institute; projects judged on responsible AI use
Notes: Pre-registration open. Evaluated on creativity, problem-solving, responsible AI use, and teamwork.

19. UCL All-Girls Hackathon — Responsible AI for Disability

Name: UCL All-Girls Hackathon 2026
Organiser: UCL Computer Science
Domain/Category: Responsible AI / Accessibility / AI Ethics
Prize Pool: Not specified
Dates: March 2026
Status: Recently ended
Format: In-person
Location: UCL, London, UK
Platform: UCL
URL: https://www.ucl.ac.uk/engineering/news/2026/mar/all-girls-hackathon-returns-focus-responsible-ai
Key Themes/Tracks: Designing responsible AI tools for people with disabilities
Notable Sponsors: UCL
Team Size: Not specified
Open to: Female-identifying students
Quality Signal: MEDIUM-HIGH — UCL is a top UK university; second edition
Notes: Focus on accessibility and responsible AI design.

RELATED RESEARCH PROGRAMS & FELLOWSHIPS (Hackathon-Adjacent)

These are not hackathons per se but are significant structured programs in the AI safety/alignment/governance space happening around the same timeframe.

20. ControlConf Berkeley 2026

Name: ControlConf 2026
Organiser: Redwood Research + FAR.AI
Domain/Category: AI Control / AI Safety Conference
Dates: April 17 (workshop) + April 18-19 (conference), 2026
Status: Upcoming
Format: In-person
Location: Berkeley, CA
URL: https://controlconf.org/
Notes: Two-day conference on AI control. Workshop on April 17 covers AI futurism and threat modeling. Feeds directly from the AI Control Hackathon results.

21. MATS (ML Alignment & Theory Scholars) — Summer 2026

Name: MATS Summer 2026
Organiser: MATS Program
Domain/Category: AI Alignment / AI Safety Research
Dates: June - August 2026
Status: Upcoming (applications forthcoming)
Format: In-person
Location: Berkeley, CA
URL: https://www.matsprogram.org/program/summer-2026
Stipend: Provided (amount not specified)
Notes: Largest MATS program to date: 120 fellows, 100 mentors. Mentors from Anthropic Alignment Science, UK AISI, Redwood Research, ARC, LawZero. Autumn 2026 applications open late April.

22. ARENA 8.0 — AI Safety ML Bootcamp

Name: ARENA 8.0
Organiser: ARENA (AI Safety Education)
Domain/Category: AI Safety / ML Engineering
Dates: May 25 - June 26, 2026
Status: Upcoming (applications closed March 8)
Format: In-person
Location: LISA (London Initiative for Safe AI), London, UK
URL: https://www.arena.education/
Support: Covers travel, visa, accommodation, meals
Notes: 4-5 week ML bootcamp with AI safety focus. Alumni have gone to MATS, LASR, UK AISI, Apollo Research.

23. LASR Labs Summer 2026

Name: LASR Labs Summer 2026
Organiser: London AI Safety Research Labs
Domain/Category: AI Safety Research
Dates: July - October 2026 (13 weeks)
Status: Upcoming (applications open until March 30)
Format: In-person
Location: LISA, London, UK
URL: https://www.lasrlabs.org/
Stipend: GBP 11,000 + food, office space, travel
Notes: Teams of 3-4 write and submit AI safety research papers. 90% of alumni work in AI safety/security (UK AISI, Apollo, OpenAI, etc).

24. Anthropic Fellows Program — May & July 2026

Name: Anthropic AI Safety Fellows
Organiser: Anthropic
Domain/Category: AI Safety / Alignment Research
Dates: Cohorts starting May 2026 and July 2026 (4 months each)
Status: May cohort closed; July cohort accepting rolling applications
Format: Remote (must be in US, UK, or Canada)
URL: https://alignment.anthropic.com/2025/anthropic-fellows-program-2026/
Stipend: $3,850/week USD (or GBP 2,310 / CAD 4,300) + ~$15K/month compute
Research Areas: Scalable oversight, adversarial robustness, AI control, model organisms, mechanistic interpretability, AI security, model welfare
Notes: 80%+ of first cohort produced papers; 40%+ joined Anthropic full-time. VERY HIGH quality signal.

25. Pivotal Research Fellowship (Q1 2026)

Name: Pivotal Research Fellowship (7th cohort)
Organiser: Pivotal Research
Domain/Category: AI Safety / Biosecurity / AI Governance
Dates: February 9 - April 10, 2026
Status: Active (currently running)
Format: In-person
Location: LISA, London, UK
URL: https://www.pivotal-research.org/fellowship
Mentors: Google DeepMind, GovAI, UK AISI, Redwood Research, SecureBio, FAR AI
Notes: Focus areas: Technical Safety, Governance & Policy, Technical Governance, AIxBio. 70% of fellows from recent cohorts received extension funding.

26. GovAI Summer Fellowship 2026

Name: GovAI Summer Fellowship 2026
Organiser: Centre for the Governance of AI
Domain/Category: AI Governance / AI Policy
Dates: June 8 - August 28, 2026
Status: Upcoming (applications closed Jan 4)
Format: In-person
Location: London, UK
URL: https://www.governance.ai/post/summer-fellowship-2026-research-track
Stipend: GBP 12,000 + travel support
Notes: Research Track (independent research) and Applied Track. Mentorship from leading AI governance experts.

27. AI Safety Camp 11 (AISC)

Name: AI Safety Camp 11
Organiser: AI Safety Camp
Domain/Category: AI Safety (various projects)
Dates: January - April 2026
Status: Active (currently running)
Format: Online (part-time)
Location: N/A (remote)
URL: https://www.aisafety.camp/
Commitment: 10 hours/week + weekly meetings
Notes: 27 public projects in this edition. Seven-year track record; alumni founded 10 organizations. Projects range from technical safety to advocacy and policy.

28. SPAR Spring 2026

Name: SPAR Spring 2026 (Supervised Program for Alignment Research)
Organiser: SPAR
Domain/Category: AI Safety / AI Policy / Biosecurity (new for 2026!)
Dates: February 16 - May 16, 2026
Status: Active (currently running)
Format: Remote (part-time, 5-20 hours/week)
URL: https://sparai.org/
Details: 130+ research projects. Mentors from Google DeepMind, RAND, Apollo Research, MATS, SecureBio, UK AISI, MIRI, Goodfire, Rethink Priorities, and universities including Cambridge, Harvard, Oxford, MIT
Notes: ~50% increase in projects vs Fall 2025. First time offering biosecurity projects. Open to undergrads through professionals.

29. ERA Fellowship — Cambridge (Winter 2026)

Name: Cambridge ERA:AI Fellowship Winter 2026
Organiser: ERA Fellowship
Domain/Category: AI Safety / AI Governance / Technical AI Governance
Dates: February 2 - March 27, 2026
Status: Active (currently running)
Format: In-person
Location: Cambridge, UK
URL: https://erafellowship.org
Stipend: Prorated from GBP 34,125/year + accommodation + meals + visa support + travel
Streams: Technical AI Safety, AI Governance, Technical AI Governance
Notes: ~30-35 fellows. 8-week intensive research program with mentor support. Applications for next cohort TBD.

30. TARA (Technical Alignment Research Accelerator) — March 2026

Name: TARA Round 1 2026
Organiser: TARA
Domain/Category: AI Safety / Technical Alignment
Dates: March 7 - June 13, 2026
Status: Active (currently running)
Format: In-person Saturday sessions + independent weekly work (14 weeks, part-time)
Location: Sydney, Melbourne, Brisbane, Manila, Tokyo, Singapore
URL: https://www.taraprogram.org
Cost: Free (lunch, compute credits, study space covered)
Notes: Uses ARENA curriculum. Requires strong Python background. Covers transformers, mechanistic interpretability, RL, model evaluation. Culminates in 3-week publication-quality project.

31. Iliad Intensive (April 2026)

Name: Iliad Intensive
Organiser: Principles of Intelligence (PIBBSS)
Domain/Category: Theoretical AI Alignment
Dates: April 4 - May 2, 2026 (also June and August iterations)
Status: Upcoming (application deadline was March 6)
Format: In-person
Location: London, UK
URL: https://www.iliad.ac/
Notes: Month-long intensive on theoretical alignment: RL theory, learning theory, interpretability, agent foundations, scalable oversight. Selects for technical excellence in math, theoretical physics, theoretical CS. Can lead to Iliad Fellowship.

32. XLab Summer Research Fellowship 2026

Name: XLab Summer Research Fellowship
Organiser: Existential Risk Laboratory, University of Chicago
Domain/Category: AI Safety / Nuclear Security / Existential Risk
Dates: Summer 2026 (10 weeks)
Status: Upcoming (application deadline March 15)
Format: In-person
Location: University of Chicago campus
URL: https://xrisk.uchicago.edu/srf-26/
Stipend: $7,500 + housing/meals OR $9,500 (off-campus)
Notes: 15-20 fellows. Workshops on forecasting and OSINT. Cohort-based with peer review sessions.

33. Second Look Fellowship 2026

Name: Second Look Fellowship
Organiser: XLab, University of Chicago
Domain/Category: AI Safety Research / Replication
Dates: June 15 - August 22, 2026
Status: Upcoming (rolling applications until March 15)
Format: In-person
Location: Chicago, USA
URL: https://secondlookresearch.com/fellowship
Notes: Unique focus: producing high-quality replications of influential AI safety research. 10-week program.

34. CAIS AI and Society Fellowship 2026

Name: AI and Society Fellowship
Organiser: Center for AI Safety
Domain/Category: AI Safety / AI Policy / Social Impact
Dates: June 1 - August 21, 2026
Status: Upcoming
Format: In-person
Location: San Francisco, CA
URL: https://safe.ai/fellowship
Notes: Provides exposure to frontier AI developments with peer community focused on AI as central issue.

35. CLR Summer Research Fellowship 2026

Name: CLR Summer Research Fellowship 2026
Organiser: Center on Long-Term Risk
Domain/Category: AI Safety / S-Risk / Suffering Reduction
Dates: Summer 2026 (exact dates TBD)
Status: Upcoming
Format: Not specified
URL: https://longtermrisk.org/summer-research-fellowship/
Notes: Focus on s-risk relevant empirical AI safety work. Model Persona research agenda — understanding when AI personas develop malicious traits (spitefulness, sadism, punitiveness).

36. Algoverse AI Safety Fellowship (Spring 2026)

Name: Algoverse AI Safety Research Fellowship
Organiser: Algoverse AI Research
Domain/Category: AI Safety Research
Dates: January 26 - May 1, 2026
Status: Active (currently running)
Format: Online (part-time, 12 weeks)
URL: https://algoverseairesearch.org/ai-safety-fellowship
Cost: Free (funded by Open Philanthropy grant)
Notes: 30 fellows selected from 60 initial participants. Expert-led sessions + hands-on projects. Open to university students and industry professionals worldwide.

37. CHAI Workshop 2026

Name: 10th Annual CHAI Workshop
Organiser: Center for Human-Compatible AI, UC Berkeley
Domain/Category: AI Safety / Human-Compatible AI
Dates: June 4-7, 2026
Status: Upcoming
Format: In-person (retreat-style)
Location: Asilomar Hotel & Conference Grounds, Pacific Grove, CA
URL: https://workshop.humancompatible.ai/
Notes: ~250 researchers, practitioners, policymakers. Poster session with abstract submissions. Chatham House Rule. 10th anniversary edition.

38. Effective Impact Hackathon (EA Sweden)

Name: Effective Impact Hackathon
Organiser: EA Sweden, EA KTH, EA Uppsala, EA SASSE
Domain/Category: Effective Altruism / Social Impact
Prize Pool: Not specified
Dates: 2026 (specific dates TBD — recurring event)
Status: Upcoming/Recurring
Format: In-person
Location: Sweden
URL: https://www.effektivaltruism.org/en/effective-impact-hackathon/
Team Size: 3-4 persons
Notes: Teams present 5-minute pitches to jury. EA-aligned hackathon.

39. AI Security Bootcamp: Singapore 2026

Name: AI Security Bootcamp: Singapore 2026
Organiser: Partnership with Singapore AI Safety Hub (SASH)
Domain/Category: AI Security / AI Safety
Dates: April 20-26, 2026
Status: Upcoming
Format: In-person (intensive 1-week bootcamp)
Location: Singapore
Notes: Trains security professionals to secure frontier AI systems.

40. PauseCon DC 2026

Name: PauseCon DC 2026
Organiser: PauseAI / AI safety advocacy community
Domain/Category: AI Safety Advocacy / Policy
Dates: April 11-14, 2026
Status: Upcoming
Format: In-person
Location: Washington D.C., USA
Notes: 4-day conference with workshops, grassroots lobbying training, meetings with congressional offices. Not a hackathon but a significant AI safety community event.

SUMMARY TABLE — Hackathons & Competitions Only (Excludes Fellowships/Programs)

#	Name	Prize Pool	Dates	Status	Format	Quality
1	AI Control Hackathon 2026	$2K + funded trip	Mar 20-22	ACTIVE	Hybrid	HIGH
2	AI Manipulation Hackathon	$2K	Jan 9-11	Ended	Hybrid	HIGH
3	Technical AI Governance Challenge	~$2K	Jan 30-Feb 1	Ended	Hybrid	HIGH
4	HackAPrompt 2.0	$100K	Ongoing	Active	Online	HIGH
5	OpenAI Red-Teaming Challenge	$500K	Aug 2025	Ended	Online	VERY HIGH
6	SafeBench	$250K	Mar 2024-Feb 2025	Ended	Online	VERY HIGH
7	IEEE SaTML 2026 Competitions	Academic	Mar 23-25	Active	Hybrid	VERY HIGH
8	Amazon Nova AI Challenge	~$3.1M+	2026 ongoing	Active	Hybrid	VERY HIGH
9	Mila Youth Mental Health Hack	CAD 10K	Mar 16-23	Active	Hybrid	HIGH
10	CIVICA EU AI Hack	TBD	Apr 17-19	Upcoming	In-person	MEDIUM-HIGH
11	iSAFE Hackathon 2026	Cash+credits	Feb 8-9	Ended	Hybrid	MEDIUM-HIGH
12	Society-Centered AI (Duke)	Prestige	Feb 12-14	Ended	In-person	HIGH
13	AI Safety Fieldbuilding Hack	TBD	Mar 28	Upcoming	In-person	MEDIUM
14	Re-Align Challenge (ICLR)	Academic	Apr 26-27	Upcoming	Hybrid	HIGH
15	IEEE IES GenAI Challenge	Academic	Feb-Apr	Active	Online	HIGH
16	ISB AI Safety Hackathon	TBD	2026 TBD	Upcoming	TBD	MEDIUM
17	CISPA EU Hackathon Championship	TBD	2026 ongoing	Active	In-person	HIGH
18	USAII Global AI Hackathon	$15K	Jun 2026	Upcoming	Online	MEDIUM
19	UCL All-Girls Hackathon	TBD	Mar 2026	Ended	In-person	MEDIUM-HIGH