Guidelines for Artificial Intelligence Containment
This work addresses safety concerns for the AI research community, but it appears incremental as it builds on previous work on the AI Containment Problem.
The paper tackles the need for safety software in AI research by proposing guidelines for developing reliable sandboxing software to contain intelligent programs, aiming to enable safe study and analysis of AI agents while preventing risks like information leakage and cyberattacks.
With almost daily improvements in capabilities of artificial intelligence it is more important than ever to develop safety software for use by the AI research community. Building on our previous work on AI Containment Problem we propose a number of guidelines which should help AI safety researchers to develop reliable sandboxing software for intelligent programs of all levels. Such safety container software will make it possible to study and analyze intelligent artificial agent while maintaining certain level of safety against information leakage, social engineering attacks and cyberattacks from within the container.