President Joe Biden issued a wide-ranging executive order on Monday to place new regulations on the development of artificial intelligence technology. The new policies implement additional oversight for emerging AI models, introducing user protections and laying a foundation for future legislation and global regulations.
“AI is all around us,” Biden said at a signing event on Monday afternoon. “To realize the promise of AI and avoid the risk, we need to govern this technology.”
Using the powers of the Defense Production Act, the Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence requires leading AI developers—including Microsoft, Amazon, and Alphabet/Google—to keep the government informed about safety test results and other developments.
The order also instructs the National Institute of Standards and Technology to create public safety standards for new AI tools before release. The Department of Energy is tasked with ensuring that AI cannot be used to develop chemical, biological, or nuclear weapons, the Departments of Defense and Homeland Security will safeguard digital infrastructure, and the Department of Commerce is to develop a process for watermarking AI-generated content to prevent misinformation.
Biden remarked on the latest advances in “deepfake technology,” remarking that he had watched AI-generated videos of himself and was left wondering “When the hell did I say that?”
“AI devices are being used to deceive people,” said the President.
The executive order also touches on issues of privacy, consumer protection, and workers’ rights, as well as civil rights issues like “algorithmic discrimination” to ensure that “AI advances equity.”
To accomplish these goals, the Biden administration is leveraging the US government’s position as a leading customer of Big Tech companies, encouraging AI developers to deliver the safety features that most have already promised voluntarily.
Many of the policies established in Biden’s order will need to be augmented by Congressional legislation. To that end, the president is scheduled to meet with Senate Majority Leader Chuck Schumer and a bipartisan group of lawmakers on Tuesday to further develop plans for privacy protections.
The regulations of the Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence are to be implemented between 90 and 365 days after the signing.
Add comment