Global Experts Unite For First International AI Safety Report

International experts, including representatives from 30 nations, the EU, and the UN, are joining forces to create the first International Scientific Report on Advanced AI Safety.

The aim is to create a collective understanding of the safety of advanced AI systems by amalgamating insights from scientific literature covering their capabilities and risks.

The Expert Advisory Panel, comprising 32 experts, has a key role in advising on the report’s development. Some noteworthy members are Dr. Hiroaki Kitano, CTO of Sony; Amandeep Gill, UN Tech Envoy; and Professor Yoshua Bengio, acknowledged as one of the pioneers of AI, leading the panel.

Technology Secretary Michelle Donelan spoke about diverse opinions, ensuring a global discussion on the safe development of AI. The decision to rename the report as the International Scientific Report on Advanced AI Safety reinforces its global focus, aligning with the Bletchley Declaration.


Transparent Principles And Procedures


The report’s guidelines and steps stress the need to be clear and inclusive. Reports should be open, cover everything, and be based on thorough evaluations of scientific, technical, and socio-economic evidence.

The process includes an Expert Advisory Panel, a Writing Group, and a Secretariat.

The Chair, Professor Yoshua Bengio, has ultimate responsibility, with the Expert Advisory Panel providing advice and the Writing Group drafting the report.

The Secretariat oversees administrative tasks. Importantly, the report’s neutrality on policy issues ensures a focus on objective scientific assessment.

The timeline reveals a commitment to timeliness and adaptability. The first report, scheduled for Q2 2024, will undergo a second iteration in Q4 2024.

Plans for post-2024 reports are being explored, showing an evolving approach. These principles and procedures will undergo regular reviews and updates to align with the dynamic nature of AI safety discussions.


A Diverse And Influential Advisory Panel


The Expert Advisory Panel includes representatives from diverse nations, fostering a global perspective. Members, nominated by their respective countries, contribute expertise from leading AI nations.

Notable figures include Professor Bronwyn Fox from Australia, Professor Antonio Krüger from Germany, and Saif M. Khan from the U.S. Department of Commerce. Each member will play a vital role in advising on the content of the report, reflecting the international collaboration underpinning AI safety discussions.


Building Momentum: International Collaboration In Action


The press release reveals that the international collaboration is gaining momentum. The report’s initial findings are set to be released ahead of the Republic of Korea’s AI Safety Summit in the Spring, with a second publication coinciding with talks hosted by France.

This strategic timing aligns with the broader effort to inform discussions at future AI Safety Summits, shaping the global narrative on AI safety.

Secretary of State for Science, Innovation, and Technology, Michelle Donelan, expressed enthusiasm, stating, “The International Scientific Report on Advanced AI Safety will be a landmark publication, bringing the best scientific research on the risks and capabilities of frontier AI development under one roof.” Professor Yoshua Bengio spoke on the importance of international talent contributing to the report, specifically, its importance in informing discussions at upcoming AI Safety Summits.


Country Contributions: A Global Tapestry Of Expertise


The breakdown of the Expert Advisory Panel members by country showcases a rich tapestry of global expertise. Representatives from diverse nations, such as Canada, China, India, and the UK, bring unique perspectives to the table.

The nomination of individuals like Professor Balaraman Ravindran from India, Dr. Yi Zeng from China, and Dame Angela McLean from the UK demonstrates the commitment of nations to contributing to the global dialogue on AI safety.

The comprehensive list of panel members from 30 leading AI nations, along with the EU and UN, is reflective of the inclusive nature of the initiative. Each member represents not just their country but also contributes to the broader international effort to ensure AI development aligns with global safety standards.