DARK SECRETS: 2 | Vintage Vibes

Google DeepMind agrees to sweeping research collaboration with the U.K. government

The collaboration will see the AI company collaborating with the British government on a robotic lab for new materials, fusion energy, and new research into AI safety and the societal impacts of AI

Fortune · Dec 11, 2025

AI will ‘pain whole swathe of UK economy before providing net benefit’

The greatest AI-driven cyber threat is its ability to find weak spots in digital spaces, the CTO of the National Cyber Security Centre has said.

Evening Standard · Oct 9, 2025

Mitigating AI-related risks: soft approach, hard approach or something in the middle?

As AI technology continues to evolve, the differences between regulatory approaches is becoming more apparent.

TechRadar · Apr 15, 2025

UK government’s AI plan gives a glimpse of how it plans to regulate the technology

The UK government recently published its plan for using AI to boost growth and deliver services more efficiently. It also suggests a fundamental shift in how the UK aims to position itself as a global leader in AI innovation.

The Conversation · Jan 24, 2025

British AI startup with government ties is developing tech for military drones

Exclusive: Concerns raised over role of Faculty AI, which has worked with NHS and government safety body

The Guardian - UK · Jan 7, 2025

Claude 3.5 suggests AI’s looming ubiquity could be a good thing

In this week’s newsletter: If you don’t like chatbots popping up everywhere, get ready to be peeved. But the latest version of Anthropic’s shows AI is becoming more useful – and, crucially, affordable

The Guardian - UK · Jun 25, 2024

UK AI Safety Institute to open new US office

Britain expands its AI Safety Institute by opening a new office in the US, as well as committing to Canada collaboration.

TechRadar · May 20, 2024

AI guardrails can be easily beaten, even if you don't mean to

Generative AI chatbots like ChatGPT and Gemini can be jailbroken to give malicious answers way too easily.

TechRadar · May 20, 2024

AI chatbots’ safeguards can be easily bypassed, say UK researchers

Five systems tested were found to be ‘highly vulnerable’ to attempts to elicit harmful responses

The Guardian - UK · May 19, 2024

AI safeguards can easily be broken, UK Safety Institute finds

Researchers find large language models, which power chatbots, can deceive human users and help spread disinformation

The Guardian - UK · Feb 9, 2024