How Should We Regulate AI?
654 Downloads
Autonomous and intelligent systems (A/IS) are being increasingly deployed for the purpose of making consequential decisions that impact the lives, liberty, and well-being of individuals in areas such as hiring, housing, credit, and criminal justice. There is growing concern about generation of disinformation and toxic content, threats to critical infrastructure, and even existential risks to humanity. Policy makers around the world are debating regulatory approaches to control automated systems, especially in response to growing concern over generative AI technologies like ChatGPT and DALL-E. This article shows how and why we should apply the practical and time-tested strategies for risk management from IEEE Standard 1012 for System, Software, and Hardware Verification and Validation1 (IEEE 1012) to the current hotly-debated conversation about AI regulation.
Size: 4.9 MB
How Should We Regulate AI?
654 Downloads
Autonomous and intelligent systems (A/IS) are being increasingly deployed for the purpose of making consequential decisions that impact the lives, liberty, and well-being of individuals in areas such as hiring, housing, credit, and criminal justice. There is growing concern about generation of disinformation and toxic content, threats to critical infrastructure, and even existential risks to humanity. Policy makers around the world are debating regulatory approaches to control automated systems, especially in response to growing concern over generative AI technologies like ChatGPT and DALL-E. This article shows how and why we should apply the practical and time-tested strategies for risk management from IEEE Standard 1012 for System, Software, and Hardware Verification and Validation1 (IEEE 1012) to the current hotly-debated conversation about AI regulation.
Size: 4.9 MB