Our Work
Three programmes, one purpose.
Everything the Foundation does falls into three connected areas of work — each chosen because it addresses a specific gap between AI's capabilities and the public's ability to understand, govern, and benefit from it.
Independent Research
We commission, conduct, and publish research on how artificial intelligence systems actually behave in the world — beyond benchmarks and marketing claims. Our work examines reliability, bias, transparency, and the real conditions under which these systems succeed or fail.
Because we accept no industry funding, our findings carry no obligation to flatter a sponsor. We publish what the evidence shows, including when it is inconvenient or inconclusive.
- Empirical studies of AI system behaviour and failure modes
- Independent evaluation of claims made about AI capabilities
- Analysis of fairness, bias, and disparate impact in deployed systems
- Research on transparency, explainability, and meaningful human oversight
- Working papers and briefings for policymakers and practitioners
- Plain-language explainers on how modern AI works
- Curriculum resources for secondary and higher education
- Briefings designed for journalists and newsroom editors
- Public talks, workshops, and open-access reference material
- Guidance for professionals encountering AI in their fields
Public Education & Literacy
A democratic society cannot govern what it does not understand. Our education programme produces clear, accurate, freely available resources that help people reason about artificial intelligence — what it can do, what it cannot, and what questions to ask.
We write for the thoughtful non-expert: the student, the clinician, the teacher, the reporter, the public servant. No coding background required, and no hype or fear-mongering supplied.
Start with the Journal →Responsible Deployment Standards
Principles like fairness, transparency, and accountability are easy to endorse and hard to implement. Our standards work translates them into concrete, adoptable practice — the checklists, review processes, and documentation templates that let an organisation deploy AI responsibly rather than merely promise to.
We develop this guidance in the open and offer it freely, with particular attention to high-stakes settings such as healthcare, education, and public services.
- Practical readiness assessments for AI adoption
- Model and data documentation templates
- Human-oversight and escalation design patterns
- Procurement guidance for institutions buying AI systems
- Sector-specific guidance for healthcare and education
Collaboration
We work with universities, public institutions, and civil society.
If your institution is grappling with how to study, teach, or responsibly deploy artificial intelligence, we would like to hear from you.
Start a conversation →