International Scientific Report on Advanced AI Safety: expert advisory panel and principles and procedures
How an international expert advisory panel will advise on the International AI Safety Report using a set of principles and procedures.
Documents
Details
The International Scientific Report on Advanced AI Safety will drive an up-to-date, science-based understanding of the safety of advanced AI systems. Announced at the AI Safety Summit in November 2023, the UK has commissioned Yoshua Bengio, a Turing Award-winning AI academic and member of the UN’s Scientific Advisory Board, to Chair the Report.
The report will be published in two iterations:
-
First publication in Spring 2024
-
Second publication towards the end of 2024
The second publication will build on the content of the first.
The report’s new international Expert Advisory Panel will advise on the report’s development. The Panel is made up of 30 countries including the UK and nominees from nations who were invited to the AI Safety Summit at Bletchley Park, as well as representatives of the European Union and the United Nations.
The principles and procedures document outlines how the report will ensure:
- openness of process
- transparency
- inclusivity of diverse expertise
- evidence-based evaluation
It states that the report should acknowledge where and why there is disagreement in the wider expert community, and present the debate in an objective manner.
It also outlines the need for collaborative efforts to create a consensus. In cases of ongoing differences, dissenting views can be noted in the report, promoting transparency and a diversity of perspectives.