UK's AI Security Institute appoints new director and chief strategy officer

The UK's AI Security Institute has appointed Henry de Zoete as its new director and Nate Burnikell as chief strategy officer. De Zoete is a long-term AI advisor to the government, while Burnikell joins from a background in strategy and special projects.

Meanwhile, Edward Foley, a seasoned expert in AI infrastructure security, has joined GovAI as a Research Scholar. Foley previously worked at the UK AI Security Institute, where he led strategy and special projects.

Chinese lab DeepSeek has released a new multimodal model called V4-Flash-Vision-Exp, designed for specific use cases. The model can handle visual understanding and tool use, making it suitable for agent workflows.

A new benchmark has been conducted to evaluate AI antibody design tools, revealing a significant gap between in silico confidence and experimental performance. This highlights the need for more rigorous validation of AI tools in drug discovery.

AI is widely used to create and maintain medical records, according to a study. While there are benefits such as saving time, there are concerns about errors and lack of oversight.

Rural hospitals face significant challenges in implementing AI systems, including high costs, limited resources, and lack of expertise. A new report recommends starting with measurable, lower-risk use cases.

Connecticut has enacted an omnibus AI law, representing one of the most wide-ranging attempts to regulate AI in the US. The law has implications for businesses and individuals alike.

Sen. Ed Markey and Rep. Seth Moulton have called for AI safeguards in their final debate, emphasizing the need for national laws and regulations to address the risks associated with AI.

Taiwan and Japan are collaborating on a new semiconductor ecosystem, aiming to create a new paradigm for the semiconductor industry.

Experts discuss the opportunities and challenges of agentic AI, enabling systems to reason, plan, and use tools on their own. A secret Hollywood group has also developed a framework for AI in the entertainment industry.

Key Takeaways

• DeepSeek releases V4-Flash-Vision-Exp multimodal model for specific use cases.
• UK's AI Security Institute appoints Henry de Zoete as director and Nate Burnikell as chief strategy officer.
• Edward Foley joins GovAI as Research Scholar, focusing on AI governance research.
• AI antibody design benchmark reveals validation gap between in silico confidence and experimental performance.
• AI widely used in medical records, but concerns about errors and oversight remain.
• Rural hospitals face challenges in implementing AI systems, including high costs and limited resources.
• Connecticut enacts omnibus AI law, regulating AI in the US.
• Sen. Ed Markey and Rep. Seth Moulton call for AI safeguards and national regulations.
• Taiwan and Japan collaborate on a new semiconductor ecosystem.
• Experts discuss agentic AI and its opportunities and challenges.

UK's AI Security Institute Appoints New Director and Chief Strategy Officer

The UK government’s AI Security Institute has appointed a new director and chief strategy officer to its leadership team. Henry de Zoete, a long-term AI advisor to the government, has been named as the director of the AI Security Institute. Nate Burnikell has been appointed as the organisation’s chief strategy officer. The appointments follow the departure of interim director Adam Beaumont, who is returning to a previous position at GCHQ.

Edward Foley Joins GovAI as Research Scholar

Edward Foley, a seasoned expert in AI infrastructure security, has joined GovAI as a Research Scholar. Foley previously worked at the UK AI Security Institute, where he led strategy and special projects. He will focus on AI governance research at GovAI, building on his experience in frontier AI partnerships and AI infrastructure security.

DeepSeek Releases V4-Flash-Vision-Exp Multimodal Model

Chinese lab DeepSeek has released a new multimodal model called V4-Flash-Vision-Exp. The model is designed for specific use cases and has achieved comparable performance to Opus 4.8 on some benchmarks. DeepSeek's model can handle visual understanding and tool use, making it suitable for agent workflows.

AI Antibody Design Benchmark Reveals Validation Gap

A new benchmark has been conducted to evaluate AI antibody design tools. The results show a significant gap between in silico confidence and experimental performance. The benchmark highlights the need for more rigorous validation of AI tools in drug discovery.

AI Commonly Used to Create and Maintain Medical Records

AI is widely used to create and maintain medical records, according to a study. The technology is often used via smartphone apps, and doctors have noted benefits such as saving time. However, there are concerns about errors and lack of oversight.

Rural Hospitals Face Challenges in Allocating AI Investments

Rural hospitals face significant challenges in implementing AI systems, including high costs, limited resources, and lack of expertise. A new report recommends that rural hospitals start with measurable, lower-risk use cases and address broadband, workforce, data, and governance gaps.

Connecticut Enacts Omnibus AI Law

Connecticut has enacted an omnibus AI law that represents one of the most wide-ranging attempts to regulate AI in the US. The law has implications for businesses and individuals alike.

Markey and Moulton Call for AI Safeguards

Sen. Ed Markey and Rep. Seth Moulton have called for AI safeguards in their final debate. They emphasized the need for national laws and regulations to address the risks associated with AI.

Taiwan and Japan Need Each Other in the AI Age

Taiwan and Japan are collaborating on a new semiconductor ecosystem. The partnership aims to create a new paradigm for the semiconductor industry, combining the strengths of both countries.

Agentic AI: The Next Phase in AI Evolution

Agentic AI represents the next phase in AI evolution, enabling systems to reason, plan, and use tools on their own. Experts discuss the opportunities and challenges of agentic AI.

Hollywood Group Develops AI Framework

A secret Hollywood group has developed a framework for AI in the entertainment industry. The framework focuses on defining how AI is used and ensuring human creative control.