Robustifying Alignment Techniques to Survive RLVR and Continual Learning

We want to measure & prevent drift in model alignment characteristics during RLVR and continual learning.

Hong Kong Hub for Apart Research's AI Collusion Sprint

AISHK will run an in‑person weekend sprint hub for elite finance talent and host the winning team during the Apart Fellowship.

AI Kurd - AI Safety & Alignment Education

Scaling open-source AI safety content, localized research breakdowns, and web tools for the Kurdish community specially and world generally

AI Safety x Content Fellowship

A three-week virtual fellowship where young Africans learn about AI safety and turn big ideas and catastrophe scenarios into videos the public can understand.

Fund Algon's Writing

Mycelium: AI alignment for nonhuman welfare

Seen & WithMe

Keep AI Whistleblowers Alive

The Enduring Path to Alignment: Inorganic Existential Social Theory

Sentient Futures Residency

Animal Welfare Midtraining Data Creation

Defining dangerous DNA with the Sequence Biosecurity Risk Consortium (SBRC)

Defining which sequences are dangerous enough to screen for, so providers and regulators screen consistently

Toward a Superhuman, Open-Ended Alignment Auditor

AI Safety Hong Kong (AISHK)

Save the world from deceptive AI "personas"!

To investigate how AI models form, withhold and communicate internal confidence, and whether this shapes deceptive personas affecting AI safety generalisation

NTU AI Safety - Financial support for organizers

Preserving the Human Veto

Civil-society infrastructure against AI-enabled power concentration, built around autonomous weapons

Prediction of Inoculation Prompt Side Effects

An open, cheap method that detects when an inoculation prompt inoculates against off-target traits, so labs and developers can catch undesired trait/persona cha

WithMe Test Case 001

Help Sue Private Intelligence Firms Hitting AI Whistleblowers

Sentient Futures Project Incubator

Funding compute/API costs for Incubator projects that build nonhuman welfare consideration into AI safety work

Surrogate base model for Mechanistic Interpretability

Creating a reference model for mechanistic interpretability without assuming that at auditing time we have a safe model to compare the suspicious model against.

Sydney AI Safety Space FY27

An additional year funding for the co-working hub in Sydney, Australia, which offers free office space for people working in the field of AI safety.