ai-uk-government-policy

AI Security Institute – Frontier AI Trends Report

Progress in the cyber domain is significant. Al systems are just beginning to complete expert-level cyber tasks typically requiring 10+ years of experience. Two years ago, they could barely complete tasks requiring one year of cyber expertise.

https://www.aisi.gov.uk/frontier-ai-trends-report

The quote above from this inaugural report by the UK AI Security Institute, demonstrates the rapid rate of progress currently being made by the leading AI system developers. The report outlines the capabilities of the current AI systems that are available and attempts to evaluate what risks they represent. The following excerpts were considered to be of primary interest;

“A human expert baseline was established with PhD holders in relevant biology or chemistry topics. The QA evaluations are designed to be difficult, with absolute scores for PhD holders ranging from approximately 40-50%. Even so, we’ve seen rapid progression in models’ performance up to and beyond this PhD baseline.”

“People are increasingly turning to Al systems for emotional support or social interaction. While many users report positive experiences, recent high-profile cases of harm30 underline the need for research into this area, including the conditions under which harm could occur, and the safeguards that could enable beneficial use.”

AI-UK Government Policy

Lobbying for a Pro-Worker AI Strategy

* Infographic generated using NotebookLM

AI-UK Government Policy

Jade Leung: Architecting Safety at the Edge, CTO of the UK AI Safety Institute

* Post generated using GeminiAI

As the race toward frontier AI models accelerates, the need for rigorous, independent evaluation has never been more critical. At the centre of this technical challenge in the UK sits the Artificial Intelligence Safety Institute (AISI), and steering its technical strategy is Chief Technology Officer, Jade Leung.

The role of CTO at a national safety institute is unique. It isn’t about building a commercial product; it is about building the rigorous scientific infrastructure needed to understand and mitigate the risks of superintelligent systems.

Jade Leung brings a vital dual perspective to this immense challenge. With a background spanning engineering, research at Oxford’s Future of Humanity Institute, and a prominent role leading governance research at OpenAI, she understands both the raw mechanics of LLMs and the complex sociotechnical systems they inhabit.

As CTO of the AISI, Jade’s mandate is formidable. Her key responsibilities revolve around three critical pillars:

1. Designing the “Crash Tests” for Frontier AI Jade leads the teams responsible for developing technical evaluations (“evals”) for the world’s most advanced models. This involves creating complex methodologies to stress-test models for catastrophic risks—including cybersecurity vulnerabilities, chemical and biological misuse potential, and autonomous capabilities. Her teams are essentially trying to “break” these models in a controlled environment before they are widely deployed.

2. Building Secure Research Infrastructure Testing frontier models requires immense compute power and highly secure environments. Jade oversees the engineering effort to build and maintain the technical infrastructure that allows the AISI to handle sensitive model weights and conduct large-scale experiments securely, independent of private labs.

3. Bridging the Technical-Policy Divide Perhaps most importantly, Jade’s background allows her to translate dense technical findings into actionable insights for the government. She ensures that regulatory decisions are grounded in empirical engineering reality, not just theory.

Under Jade Leung’s technical leadership, the AISI is moving beyond high-level discussions of safety and into the realm of concrete, engineering-led evaluation. It is pioneering work that is defining how nations interact with advanced artificial intelligence.

#AI #AISafety #TechLeadership #UKGovernment #FrontierAI #Engineering

AI-UK Government Policy