UK House of Lords Launches Inquiry into Large Language Models - Citylife Cape Town

2023-09-08 05:16 (EST) - Vicky Stavropoulou

The House of Lords Communications and Digital Committee in the UK is set to begin an inquiry into large language models (LLMs). LLMs are a type of generative artificial intelligence (AI) that have advanced significantly in recent years. The inquiry aims to understand the unique characteristics of LLMs and assess the opportunities and risks they present over the next three years.

The committee will hear evidence from key figures in the AI sector, including Ian Hogarth, chair of the government’s AI Foundation Model Taskforce, Jean Innes, CEO of The Alan Turing Institute, Professor Neil Lawrence from DeepMind, and Ben Brooks from Stability AI.

During the evidence session, the committee will explore how LLMs differ from other forms of AI and how they are likely to evolve in the future. They will also discuss the role and structure of the Foundation Model Taskforce, as well as the government’s role in addressing the opportunities and risks associated with LLMs.

One of the main concerns highlighted during the inquiry is the security risks posed by LLMs. Threats include prompt injections, data leakage, inadequate sandboxing, and unauthorized code execution. Ian Hogarth recently warned that AI could be exploited by cybercriminals to disrupt the National Health System (NHS) in the UK. Such attacks could have significant consequences comparable to the COVID-19 pandemic or the WannaCry ransomware attack in 2017. The National Risk Register (NRR) for 2023 has officially classified AI as a “chronic risk,” emphasizing its long-term threat as opposed to an acute one.

The inquiry aims to assess the government’s preparedness and determine priorities for action regarding LLMs. By understanding the unique nature of these AI models, the UK can effectively respond to the opportunities and risks they bring, ensuring the safety and security of critical systems.

Sources:

– The House of Lords Communications and Digital Committee to Open Inquiry into Large Language Models (source)

– AI Could Pose Threat to UK National Security and NHS, Warns AI Foundation Model Taskforce Chair (source)

Source

Previous
Previous

CISA Warning: Nation-State Hackers Exploit Fortinet and Zoho Vulnerabilities - The Hacker News

Next
Next

North Korean Hackers Target Security Researchers — Again - Dark Reading