Research
Research, without theatre.
Privacy engineering, AI agents, scientific evaluation, and UK technology governance. We show enough to prove depth; we do not publish client systems, proprietary methods, or exploitable detail.
Research programme
The public edge of the work.
Our internal research is more detailed than the website. The public layer is deliberately selective: enough to show scientific understanding, not enough to hand over competitive intelligence, evaluation prompts, exploit paths, or client-specific architecture.
AI agents and evaluation
How to measure agentic systems before they are trusted with tools, memory, workflow autonomy, or customer-facing decisions.
Privacy-enhancing technologies
Differential privacy, selective disclosure, zero-knowledge proofs, secure computation, confidential computing, and where each primitive is actually useful.
Digital identity and age assurance
The technical difference between proving an attribute and building linkable identity infrastructure.
Scientific method for product claims
Borrowing from Turing, Popper, and modern AI-for-science: define the claim, test the failure mode, publish only what survives.
Science watch
Signals from top labs.
We track public work from AISI, OpenAI, Google DeepMind, the Alan Turing Institute, and adjacent research communities. We are especially interested in evaluated agents, AI-for-science loops, and the Turing line from computation to operational tests.
AISI
Frontier evaluation is becoming infrastructure
The UK AI Security Institute frames advanced-AI governance as a scientific measurement problem: capabilities, risks, mitigations, and repeatable tests.
OpenAI LifeSciBench
Scientific agents need research-grade benchmarks
Real scientific work involves evidence handling, uncertainty, experimental design, validation, and communication; not just clean quiz answers.
OpenAI + Molecule.one
Closed-loop AI science now reaches wet-lab feedback
Near-autonomous chemistry work shows the pattern we care about: propose, test, measure, replicate, and keep expert judgement in the loop.
Google DeepMind
Evolutionary search is becoming an engineering primitive
AlphaEvolve is a useful public reference point for measurable recursive improvement: generate candidates, score them, retain what works.
Alan Turing Institute
UK AI research still starts with rigorous definitions
The Turing lineage matters: computation, evidence, decidability, and operational tests before grand claims.
- 16 July 2026·9 min read·Privacy-Enhancing Technologies·Federated Learning
Federated Computing: Deploying AI on Data You Cannot Move
Federated learning, differential privacy, and trusted execution environments let organisations train and run AI on sensitive data that regulation or competition keeps locked in place. A grounded look at what has actually shipped, what the regulators say, and where the claims outrun the evidence.
Read article - 6 May 2026·5 min read·Digital ID·UK Policy
UK Digital ID: Ten Safeguards Parliament Must Legislate
Our litigation-grade response to the Cabinet Office consultation on digital identity (CP 1498). Conditional support in principle, ten demands in primary legislation.
Read article - 8 April 2026·5 min read·Methodology·VPN
How We Score VPN Evidence: TheVPNMatrix Methodology Explained
A transparent, reproducible evidence-scoring system across 28 criteria. Every claim on TheVPNMatrix.com is backed by a record in this database.
Read article - 19 March 2026·4 min read·GDPR·Privacy
What Is Privacy-by-Design and Why It Matters for UK Businesses
Privacy-by-Design is a legal requirement under UK GDPR, not just a best practice. This guide explains the seven foundational principles and how UK businesses can implement them.
Read article - 15 March 2026·4 min read·GDPR·Startups
A Practical Guide to GDPR Compliance for UK Startups
GDPR compliance does not require a legal team or a six-figure budget. This practical guide covers the essentials every UK startup needs to know.
Read article - 10 March 2026·4 min read·AI Governance·Regulation
AI Governance: What UK Businesses Need to Know in 2026
The regulatory landscape for AI is evolving rapidly. This guide covers the EU AI Act, UK AI principles, and practical steps for building an AI governance framework.
Read article - 12 February 2026·5 min read·Facial Recognition·Civil Liberties
Facial Recognition in the UK: Built Before Parliament Authorised It
Mass biometric surveillance is being procured, installed, and operated in advance of the law that would authorise it. Our response to the Home Office consultation.
Read article