U.S. President Joe Biden has signed an executive order that establishes new standards for safety and privacy protection regarding artificial intelligence. The White House claims this step will protect Americans' information, encourage innovation and competition, and enhance the United States' leadership role in the technology field. According to CBS News, with laws lagging significantly behind technological advancements, the administration promotes the new executive order as being based on previous voluntary commitments by leading tech companies concerning safe and secure AI development. In earlier statements, the president described his executive order as "the most significant action any government has ever taken anywhere in the world regarding AI safety, security, and trust."
In a tweet on X, the White House revealed a set of rules and principles aimed at affirming America's leadership in AI regulation, including "protecting Americans' privacy, establishing equality and civil rights, and safeguarding innovation and competition."
What does the executive order mean? The order sets additional standards and requirements, as explained by CBS News. It requires AI system developers to share their safety test results with the federal government. The executive order will subject AI systems from companies like Microsoft, Google, and Amazon to oversight from various government agencies, steering the AI sector away from a path of self-regulation, according to Barron's.
This move represents "the most ambitious attempt by the U.S. government to stimulate innovation and address concerns that the flourishing technology could exacerbate bias, displace workers, and undermine national security," according to The Washington Post. The White House states that this aligns with the Defense Production Act, which mandates companies that develop models posing risks to national security, public health, or economic security to notify the federal government and share results.
The safety tests conducted by developers, known as "red team" tests, ensure that new products do not pose significant threats to users or the broader public, according to ABC News. If a product fails the safety evaluation, the federal government can compel the company either to improve the product or abandon it entirely.
The administration will also work on developing standards for biological composition testing to guard against the risky use of AI in creating hazardous biological materials. These standards will be a prerequisite for federal funding, as per CBS News.
The National Institute of Standards and Technology will establish safety standards before public release, while the Department of Homeland Security will enforce these standards in vital infrastructure sectors and create a Safety and Security Council for AI. According to the network, the Department of Energy will collaborate with the Department of Homeland Security to address threats facing infrastructure along with chemical, biological, and other risks that could arise from AI.
It is also intended that this order will enhance privacy by evaluating how agencies collect and use commercially available information, and setting guidelines for federal agencies to assess the effectiveness of privacy assurance techniques. The administration also aims to promote privacy-preserving technologies and research, such as encryption tools.
According to ABC News, one of the new standards in the executive order will regulate the use of watermarks that alert consumers when they encounter AI-supported products, which may mitigate the threat posed by deceptive content such as deepfakes.
The president's order also attempts to address what the network terms "algorithmic discrimination" so that the Department of Justice and federal civil rights offices can better investigate and prosecute civil rights violations involving AI. The administration intends to develop best practices for using AI in sentencing, bail and pretrial detention, risk assessment, crime monitoring, and forecasting, among other parts of the criminal justice system.
The executive order also promotes best practices to reduce the harms of AI while leveraging its benefits concerning employment and labor standards. The administration also seeks to capitalize on the expertise of skilled immigrants and non-immigrants with significant experience in key fields for residency, study, and work in the United States, by making the visa interview and review process more efficient.
