President Biden today issued an executive order on artificial intelligence. The order includes several key components, including: Creating new safety and security standards for AI; Protecting consumer privacy; Advancing equity and civil rights; Protecting consumers; Supporting workers; Promoting innovation and competition; Working with international partners; Developing guidance for federal agencies’ use and procurement of AI.
S. Shyam Sundar, Director of the Center for Socially Responsible Artificial Intelligence at Penn State shared his thoughts on the order:
“This is an entirely sensible approach to rein in AI while still fostering innovations in this fast-developing technology. It showcases the importance of safeguarding our security and privacy by involving all stakeholders in the process of technology development,” he said. “It calls for new standards instead of regulations. Unlike AI regulations in other countries, this executive order does not call for punishing AI developers for mishaps but still holds them accountable. This is geared toward self-governance rather than legislative governance. Some may dismiss it as toothless but this approach is probably more effective in terms of promoting new breakthroughs in AI technology and in letting the marketplace hold bad actors accountable.”
“I like that it specifically calls out the significant potential of algorithms to discriminate against certain groups of Americans in high-stakes domains such as sentencing and paroling decisions. Instead of banning the use of AI in such sensitive domains, the EO calls for coordination among federal agencies in arriving at best practices,” he continued. “AI, especially generative AI, has emergent properties, so even the developers cannot always foresee all the ways in which the technology can go wrong. Therefore, regulating it is a bit tricky and over-reaction can end up having a chilling effect. I think this EO strikes the right balance.”
You can read more of his thoughts on AI regulation in this conversation.com essay.