| Name | Type | Tags | Endorsed | Team | Raising $ |
|---|
Showing 401-444 of 444·Sorted by endorsements
| Name | Type | Tags | Endorsed | Team | Raising $ |
|---|
| Project | Research | - | - | ||
 |
| Project |
Company |
| - |
| - |
| Project | Individual | - | - |
| Project | Company | - | - |
| Project | Company | - | - |
| Project | Newsletter | - | - |
| Project | Individual | - | - |
| Project | Research | - | - |
| Project | Research | - | - |
| Project | Tooling | - | - |
| Project | Research | - | - |
| Project | Platform | - | - |
| Project | Individual | - | - |
| Project | Research | - | - |
| Project | Research | - | - |
| Project | Individual | - | - |
| Project | Tooling | - | - |
| Project | Platform | - | - |
| Project | Research | - | - |
| Project | Research | - | - |
| Project | Individual | - | - |
| Project | Comms | - | - |
| Project | Research | - | - |
| Project | Research Lab | - | - |
| Project | Comms | - | - |
| Project | Research | - | - |
| Project | Tooling | - | - |
| Project | Individual | - | - |
| Project | Media | - | - |
| Project | Individual | - | - |
| Project | Company | - | - |
| Project | Company | - | - |
| Project | Research | - | - |
| Project | Individual | - | - |
| Project | Tooling | - | - |
| Project | Company | - | - |
| Project | Education | - | - |
| Project | Research | - | - |
| Project | Individual | - | - |
| Project | Standards | - | - |
| Project | Consultancy | - | - |
| Project | Tooling | - | - |
| Project | Forecasting | - | - |
| Project | Platform | - | - |
Developing evaluation methods to determine when CV models can be trusted by addressing the gap between benchmark metrics and real-world performance.
Selma: The AI that never leaves the building.
A deterministic Rust kernel that isolates dangerous LLM agents inside a strict compiler-integrated safety sandbox.
Qualitative UX research exploring how users respond to different prototypes of a browser-based tool for identifying online misinformation.
Building reproducible, transparent, and auditable AI infrastructure for neuroscience and clinical brain health.
The Librairy is a weekly newsletter that helps individuals and groups outside the field understand and be equipped with AI safety knowledge and resources.
I've already shown that the approach is very promising. This project is about making people at AI labs and alignment academics aware of it first, then personally collaborate with them (or make others work on it, even w/o me).
A fail-closed control plane for coding-agent fleets — spend caps, audited actions, rollback — plus a public eval suite and automated red-team explorer that measure whether any control layer actually stops unsafe agent actions.
LPMs for modeling x-risk patterns across human taxonomies and agentic swarms
A no-code AI safety evaluation (including red teaming) tool that enables non-technical domain experts (e.g. social scientists, STEM experts) as well as AI engineers to design, run, and communicate effective and efficient evals.
An independent evaluation of current prompt injection defenses in large language models, producing practical recommendations for safer AI deployment.
Expanding a continuously updated database covering global panel AI data as a ground-truth benchmark which allows stress-testing frontier LLMs’ fabrication rates in global contexts (lower-income and upper-income countries’ contexts
Developing CI Theory and the Exogram framework to enable healthy, sustainable human–AI collaboration and reduce long-term AI risk.
Testing, in Washington, DC, if a AI generated messenger can build trust in communities largely unseen in AI Safety as well as a human being.
A first-person creed for AI defining what the AI is for rather than what it must not do, to be internalized during training, not tacked on after.
Four-month project to fine-tune LLMs on public value surveys, evaluate behavior via a text-based agent simulator, and run a public consultation comparing model actions to respondents’ values.
Deterministic, no-LLM-judge benchmark for how faithfully AI tracks changing beliefs. Funding v1.1: a new ambivalence metric + 20 cross-domain scenarios.
Equipping our society for people-centric AI transformation
The Veil measures how the other side of the exchange, human or artificial, changes where and how a language model computes, toward testing whether those shifts confound the activation-based tools AI safety relies on.
Researching whether many conversational AI failures share underlying behavioral dimensions — and whether those dimensions can be independently governed.
Develop a decision-theory framework for AI self-alignment in nested, cyclic, partially observable multi-agent environments where delayed punishment for welfare compromise promotes cooperative behavior.
A transmedia storytelling project using an animal characters as a metaphor for AI
a person takes an occupational Survey data (+1 or 2 fresh, third party) +a existing alignment emulation model / trigger analysis mod "what would happen" emu then do all sorts of manual tagging, for maximum real world grounding
Long-term vision is to transform the Smart Clinic Chair from a telehealth device into an AI-enhanced remote examination room capable of extending clinical intelligence into healthcare deserts, workplaces, transportation truckstops
12-week project to build and user-test a web interactive narrative prototype that communicates AI risk scenarios to non-technical young adults, producing notes, feedback data, and a public write-up.
Language model where its attention-heads are not competing but consenting. Its governors monitor the hidden state geometry and provide early-warning by predicting failure and timely modify latent processes to avoid instabilities
This architecture uses retroactive observation to identify and close system blind spots, leading to better alignment over time.
Developing practical AI oversight frameworks that keep humans in control of consequential AI decisions, starting where governance capacity is weakest.
A short film on the moral gray of AI: a single conversation between Elena, a celebrated founder whose elder-care AI harmed the people it soothed, and Joanne, the journalist who championed her and now holds her accountable.
AI that surfaces power-concentration risk in civic systems by making value trade-offs visible instead of resolving them silently.
Cantivia detects deepfakes and disinformation, and simulates how synthetic media spreads, so platforms, newsrooms and fraud teams know what's fake, who made it, and how far it'll travel.
An inline authorization platform for AI agent actions
A public, cross-disciplinary safety framework for emerging animate matter technologies
Turning published AI-safety claims into runnable, independently reproducible test packages that expose fragile assumptions, hidden failure modes, and unsupported conclusions.
A local-first control layer that keeps humans in charge of AI agents by requiring inspection, evidence, and approval before any code, file, tool, provider, browser, package, or Git action occurs.
Real-time AI voice detection system across communication channels and as an alternative to expensive and enterprise systems.
Fund experts, capacity building opportunities, and operational costs to increase the offer of courses and degrees focused on AI, AI Safety, LLMs, and Cybersecurity in Cameroon's Universities.
An open modification of Google’s AlphaTensor to search for Fast Fourier Transforms (FFTs) for the obscure Triple Product Property matrix multiplication algorithm.
Real-time dynamical hidden-state safety telemetry for language models, predicting recursive collapse from latent geometry before it manifests in outputs.
An open architecture for deterministic AI governance that separates policy enforcement from model behavior, enabling trustworthy deployment across providers.
Tools and services to help organisations comply with the EU AI Act for High-Risk AI Systems
A legal and policy decision-making platform driven by AI provides transparency and uses verifiable data to support its recommendations. It will reduce the number of times LLMs hallucinate and will allow for the safe use of LLMs in
An expert-elicited estimate of how far AI has closed the gap to superintelligence, scored by people with no financial stake in the answer, plus plain-language monitoring of what the six major labs actually change.
An online debate platform that offers cash purses to winning responses around topics covering the most prescient issues of our time.
