Group CTO & Chief AI Officer @ Cape.io · AI Safety Researcher · Cambridge AI Safety Hub & Arcadia Impact
I help boards and frontier labs turn AI into reality — safely, and at scale.
Author of Autonomous Minds (Wiley, 2025) and Grow Your Business with AI (Springer, 2023). Advisor to the Bank of England, FCA, and EU AI Act. Co-author AgentMisalignment (ICLR 2026).
live
selected work
External review of DeepMind's public scheming-inability safety case for Gemini 2.5, applying the Assurance 2.0 / Claims-Argument-Evidence methodology. Translates the review into recommendations for AI developers, AISIs and regulators on how frontier safety cases should be structured to support external scrutiny.
A propensity benchmark covering shutdown resistance, oversight evasion, sandbagging and power-seeking on frontier models. Key finding: persona system prompts shift misalignment more than the choice of model itself, with direct implications for how enterprises configure deployed agents.
How agentic AI predicts and learns to enable productivity and empowerment. Governance patterns and human-machine collaboration models for the next decade.
A practical blueprint for executives scaling AI responsibly across the enterprise — from strategy to production deployment.
speaking
about
Francisco Javier Campos Zabala has spent 25+ years building and scaling AI, data, and platform organisations across AdTech, FinTech, and SaaS — at Group CTO and CIO level inside Cape.io, Fenestra, Experian, Kantar, GroupM, and Havas. He sits at the intersection of enterprise AI and frontier safety: as Group CTO and Chief AI Officer at Cape.io he is designing one of the first fully agentic-driven creative AdTech organisations on Claude, GPT, and Codex; as a researcher at the Cambridge AI Safety Hub and the Arcadia Impact AI Governance Taskforce he co-authors work on agent misalignment and frontier safety-case review. Author of Autonomous Minds (Wiley) and Grow Your Business with AI (Springer). Advisor to the Bank of England, FCA, and the EU AI Act.
writing
Last Sunday I landed back at Heathrow after an incredibly intensive, dawn-to-dusk week in Santiago, Chile. If there is one thing I took away from this trip, it is that the Latin American market is not just preparing for the AI revolution — they are already building it. The operating rhythm was relentless. My days...
The moat of proprietary AI is officially vanishing. Open source just dropped a 1-trillion-parameter heavyweight. If you are tracking the frontier of AI, the release of Kimi K2.6 — open weights under a Modified MIT License — is a massive structural shift. With 1.1T parameters and a Mixture-of-Experts architecture, the benchmarks are staggering. K2.6 is...
Presenting complex ideas is incredibly difficult. As I pack my bags for a busy week in Chile — meeting with the team at Cape.io, our clients, and IAB Chile — I have been doing a final pass on my presentations. Whenever I do this, I always think back to the legendary MIT How to Speak...
contact
For Chief AI Officer, Head of Applied AI, NED & senior advisory conversations — or for collaborations on agentic systems and frontier safety-case review — get in touch.