The Virtue Engine

Teaching AI to understand human values through story-driven games. We're building a "moral compass" for AI that's 18% more accurate than the leading models.

TELL: The AI Detector You *Don't* Have To Trust

Training AI detectors to explain why a text is AI-generated, not just whether it is.

Resonant Systems: Scale-Invariant Ethics for AI

Implementable Measures to Replace Alignment Rules with Consonant Collaboration in a Toy Model

The AI Minute (Short-form content on AI Safety)

Gen-Z-focused multimedia project that will raise awareness of AI safety and x-risk

4-month proof of concept for curiosity-driven agents on ARC-AGI-3

Studying more human-like intelligence through constraint-aware, curiosity-driven agents on ARC-AGI-3

Verifiably-Safe Transformers

Fine-tuning a coding model to bypass the difficulty of verifying attention layers

Re-evaluating Chain-of-Thought Faithfulness

Does CoT causally drive model outputs or is it a post-hoc rationalisation? Instead of asking if CoT looks faithful, we intervene on it and observe what happens.

Expanding Francophone Awareness of AI Safety Through The Flares

This is a small grant buying a large increase in high-quality Francophone AI risk communication from a creator who has already a track record.

SEVERANT: Formally Verified, Hardware-Enforced AI Safety Architecture

Building an AI system where unsafe behaviour is physically impossible. Ethics proven in formal logic, locked in silicon, enforced at every training step.

Lyptus Research: Bridge Funding

An early-stage AI safety research group based in Sydney, Australia

Precautionary AGI Governance

Limited Legal Personhood as a Reversible Safety Instrument

AI Championship: Making AI Safety Cool for India's Middle Schoolers

Inspiring India’s Middle‑Schoolers to pursue AI Safety, Governance, and X‑Risk Work

Independent Researcher, 17, Indonesia: Foundational Theory Work on Intelligence

18+ preprints across multiple fields, all written on a 2GB RAM phone. $600 removes the only thing standing between me and the next body of work.

Emergent misalignment and memetic replication in multi-agent codebases

ClevAgent: Runtime Safety Monitoring for Autonomous AI Agents

Open source agent monitoring tools to detect failures, infinite loops, and unsafe behavior in production AI systems

Interpretable Multi-LLM Judge System for Alignment Evaluation

Auditing and improving LLM-as-a-judge systems via interpretable aggregation of preferences

Career Transition into AI Control after Postdoc

KIRI-A1: Autonomous AI Agent for Full OS Control via Vision

An advanced agent that perceives your screen and executes tasks by controlling the mouse, acting as a digital proxy to handle complex work on your behalf.

Human-AI Complementarity for Identifying Harm

Safe AI Germany (SAIGE)

Germany’s talents are critical to the global effort of reducing catastrophic risks brought by artificial intelligence.