It isn't you.
The IN Agent’s code was written by one AI, adversarially reviewed by another, and the article about it was written by the AI reviewer.
Trust is the unsquared circle at the heart of the agentic AI boom. The more useful an agent becomes, the riskier it gets to give it the access it needs to be useful. The industry is trying to paper over the widening gap with increasingly elaborate technical scaffolding but the core problem isn't just technical. It's social. It's legal. It's about judgment.
Every time data-sharing fails catastrophically, the institutional reflex is the same: centralise. Build a bigger database. Connect everything. The logic feels irresistible. If the problem is that systems don't talk to each other, surely the answer is to put all the data in one place? But that impulse is killing people.
Gartner predicts that by 2028, a quarter of enterprise breaches will trace back to agentic AI attack vectors. The problem is real, the attention is welcome, but the solutions being proposed should give every enterprise architect pause for thought.
Why
The gathering pace of agentic AI deployment is fast outstripping the structural capacity of the security frameworks meant to govern it. The gap between what it can do and how we contain it is very rapidly turning into an abyss. We believe the window for getting ahead of this is narrow and closing fast.
Our position paper explains the urgency and sets out the case for a fundamentally different approach.
What
HACCU is an independent R&D initiative focusing on the security and deployment challenges of bringing agentic AI into the mainstream. As autonomy becomes automated and acquires capacity, the attack surface increases exponentially. Growing friction around domestic and industrial adoption exposes the inadequacies of existing security architectures, and the traditional tech-first strategy of 'patch the last thing that broke' isn't aging well. There is an urgent need for searching conversations about the policies and frameworks that connect societal best practice with the nuts-and-bolts of integrating agentic AI into everyday workflows.
HACCU is investigating a profoundly different perspective. Drawing on established principles of human delegation, contractor management, and bounded authority, we are building reference implementations to test and validate practical ways of translating governance theory into effective deployment strategies.
The underlying proposition is simple. Working models for managing delegated trust already exist. They just haven't been applied to AI yet. Whoever solves this problem first won't just make agentic AI safer, they'll own the deployment landscape.
How
HACCU's research activity is focusing on identifying and articulating the structural weaknesses in current agentic AI security models.
In parallel, we are developing reference implementations in the lab - working prototypes that test how our theoretical frameworks hold up under real-world conditions. The aim is to evolve these into deployable products that organisations and individuals can use to guide their own internal policies and to integrate agentic AI safely and effectively into their existing and future workflows. An initial prototype using open-weight models for on-device deployment is in active development, with a version 1 commitment to integration with cloud-based enterprise LLMs and industry-standard project management and messaging suites.
We are also actively listening to users, industry leaders, and policy makers to ensure that what we build reflects the real-world experience and needs of the people, organisations, and institutions who depend on it. If this is your field, we'd love to hear from you.
We welcome conversations with anyone interested in supporting our work.
When
Our industry briefings are available now and already attracting attention from researchers and policy professionals. A working prototype of our reference implementation is in active development and due to enter testing Q2 2026. Further papers reporting on the build process and the practical lessons it yields are being released in parallel.
We are in early-stage conversations with technology partners and sponsors and expect to announce formal collaborations in the near future.
Where
HACCU is based in the United Kingdom and operates internationally. We welcome enquiries from researchers, policy makers, industry professionals, and from anyone who cares about getting this right.
Start the conversation - get in touch.
Who
Professor Peter McBurney
Advisory Board
Peter is Professor of Computer Science and former Head of the Department of Informatics at King's College London. He is currently Director of the NCSC Academic Centre of Excellence in Cyber Security Research at King's College London.
Peter holds a first degree and University Medal in Pure Mathematics and Statistics from the Australian National University, and a PhD in Artificial Intelligence from the University of Liverpool. His research focuses on autonomous agents, agent communication protocols, multi-agent simulation, and automated decision-making, with over 11,700 academic citations.
He is a former joint Editor-in-Chief of the academic AI journal, The Knowledge Engineering Review, and is a former member of the Academic Advisory Council of the UK Financial Conduct Authority. He has also been Co-Head of Technology Consulting at law firm Norton Rose Fulbright, advising global clients on AI, blockchain, and distributed ledger technologies. Earlier in his career, he co-founded Redwing Consulting, a consultancy company which advised leading global telecommunications companies on market entry strategies and their implementation.
His research has been funded by the UK EPSRC, Innovate UK, the EC, the ILO, the International Development Research Centre of Canada, and the Ford Foundation.
Dave Shuker
Advisory Board
Dave is a telecommunications consultant and investment advisor with deep experience in navigating the gap between emerging technology and real-world deployment. He co-founded Redwing, a global consultancy advising mobile telcos and MVNOs on market entry strategy, where he helped establish mobile carriers for Verizon, Qualcomm, Axiata and many others. He played a leading role in steering Redwing's eventual acquisition by Omnicom subsidiary Agency.com.
He subsequently founded Redwing Asia, a Singapore-based consultancy focusing on the Southeast Asian technology sector, and went on to advise on how new platforms gain traction in markets where infrastructure, regulation, and trust frameworks are still maturing. His analysis of disruptive technology adoption in emerging markets has been cited in the Financial Times and the Wall Street Journal.
Dave brings HACCU proven expertise in forecasting how emerging technologies reshape existing platforms, and in identifying the opportunities and challenges that emerge when they do.
Jim Curry
Founder
Jim is a technology entrepreneur whose career spans the emergence of the mobile web through to the current era of agentic AI. He started the mobile internet divisions at web agencies Agency.com and iXL Inc, before co-founding iFone Mobile Games where he served as CTO and played a leading role in securing start-up funding from Infogrames Entertainment. iFone was eventually acquired by Glu Mobile.
Jim brings HACCU deep experience of finding the signal in the noise, and of designing products that deliver tangible value at the boundary between emerging platforms and established industries.