A recent interim report on the safety of artificial intelligence (AI) has revealed that AI experts do not universally agree on various aspects regarding the current capabilities of AI and its potential evolution over time. The report, titled the International scientific report on the safety of advanced AI, was a significant outcome of the Bletchley Park discussions and the Bletchley Declaration.
The report explores differing opinions on the likelihood of extreme risks associated with AI that could have an impact on society, such as widespread unemployment, AI-enabled terrorism, and a loss of control over the technology. Despite these differences, the experts who contributed to the report agreed that society and policymakers need to prioritize improving their understanding of the impact of AI technology.
Yoshua Bengio, the chair of the report, emphasized the potential of AI to bring positive transformative change to various aspects of our lives when used responsibly and regulated effectively. However, due to the significant impacts, the dual use of AI, and the uncertainty surrounding its future trajectory, Bengio stressed the importance of working together to mitigate associated risks.
The report emphasizes the need for governments, academia, and society as a whole to advance the agenda of AI safety in order to harness the benefits of AI safely and successfully. The report was initially launched as the State of Science report and brings together a diverse team of global AI experts, as well as an Expert Advisory Panel from 30 AI nations, representatives from the United Nations, and the European Union.
The Secretary of State for Science, Innovation, and Technology, Michelle Donelan, expressed the importance of capturing the incredible opportunities presented by AI in a safe and responsible manner. Donelan stated that the report will inform discussions at the upcoming AI Seoul Summit and build upon the legacy of Bletchley Park by incorporating scientific evidence to advance the global conversation on AI safety.
The interim report focuses on advanced “general-purpose” AI, including systems capable of generating text and images and making automated decisions. The final report, set to be published in time for the AI Action Summit in France, will incorporate feedback from industry, civil society, and various representatives from the AI community to ensure it remains up-to-date with the latest research and provides a comprehensive view of the risks associated with advanced AI.
Marietje Schaake, International Policy Director at Stanford University Cyber Policy Center, emphasized the urgent need for democratic governance of AI based on independent research. Schaake praised the interim report for bringing together expert views on the evolution of general-purpose AI, its risks, and future implications, and stressed the importance of public leaders taking action to inform and safeguard society regarding AI, mitigating present-day harms and preparing for the consequences of more powerful AI systems in the future.