GOV UK. International Scientific Report on the Safety of Advanced AI.
Forewords
This report is the beginning of a journey on AI Safety I am honoured to be chairing the delivery of the inaugural International Scientific Report on Advanced AI Safety. I am proud to publish this interim report which is the culmination of huge efforts by many experts over the six months since the work was commissioned at the Bletchley Park AI Safety Summit in November 2023.
We know that advanced AI is developing very rapidly, and that there is considerable uncertainty over how these advanced AI systems might affect how we live and work in the future. AI has tremendous potential to change our lives for the better, but it also poses risks of harm. That is why having this thorough analysis of the available scientific literature and expert opinion is essential. The more we know, the better equipped we are to shape our collective destiny.
Our mission is clear: to drive a shared, science-based, up-to-date understanding of the safety of advanced AI, and to continue to develop that understanding over time. The report rightly highlights that there are areas of consensus among experts and also disagreements over the capabilities and risks of advanced AI, especially those expected to be developed in the future. In order to meet our mission effectively, we have aimed to address disagreement amongst the expert community with intellectual honesty. By dissecting these differences, we pave the way for informed policy-making and stimulate the research needed to help clear the fog and mitigate risks.
I am grateful to our international Expert Advisory Panel for their invaluable comments, initially shaping the report’s scope and later providing feedback on the full draft. Their diverse perspectives and careful review have broadened and strengthened this interim report. Equally deserving of recognition are my dedicated team of writers and senior advisers. Their commitment over the past few months has created an interim product that has surpassed my expectations. My thanks also go to the UK Government for starting this process and offering outstanding operational support. It was also important for me that the UK Government agreed that the scientists writing this report should have complete independence.
This interim report is only the beginning of a journey. There are no doubt perspectives and evidence that this report has failed to capture in this first attempt. In a scientific process such as this, feedback is precious. We will incorporate additional evidence and scientific viewpoints as we work toward the final version.
Professor Yoshua Bengio
Université de Montréal / Mila – Quebec AI Institute & Chair