Press Release: CDT Marks One-Year Anniversary of President Biden’s Executive Order on Artificial Intelligence
(WASHINGTON) — The Center for Democracy & Technology (CDT) today marked the one-year anniversary of President Joe Biden’s Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. CDT was a leading voice advocating for the Biden Administration to issue an executive order on AI, and has since worked closely on the EO’s implementation across federal agencies.
“The federal government has an obligation to ensure that industry practices and its own conduct don’t harm people’s rights, and that tax dollars are used effectively and responsibly. Through this order — and the work that has followed over the past year — the Administration has taken significant steps towards making that a reality when it comes to artificial intelligence,” said Alexandra Reeve Givens, president and CEO of the Center for Democracy & Technology. “While this is only a first step, in the last year we’ve seen agencies lay out important guidelines for protecting civil rights and civil liberties, as well as public safety, in the public and private sectors. The work builds on bipartisan progress that’s been made in recent years—and it rightly acknowledges that sustainable AI innovation can only be achieved with appropriate governance and guardrails.”
The AI Executive Order launched a whole-of-government effort to advance the responsible development and use of AI, including important guidance for the federal government’s own use and acquisition of the technology:
- The Office of Management and Budget (OMB) issued guidance for federal agencies’ use of AI, requiring agencies to implement and publicly document how they are mitigating any risks of AI systems that impact the safety and human and civil rights of the public.
- OMB issued further guidance about responsible government procurement of AI, directing agencies to ensure that third-party vendors implement needed safeguards for any AI product purchased by the government.
- The Administration issued a Memo and Framework on the use of AI in National Security contexts.
- The Department of Agriculture and the Department of Health and Human Services also issued guidance for state and local governments about how to identify and mitigate AI risks in public benefits programs.
CDT welcomed each of these important steps, while noting, in particular, that the National Security Memorandum and its accompanying Framework fall short because of the lack of transparency, independent oversight, and accountability they envision in an area that can significantly impact people’s rights. For example, governmental uses of AI in the national security context that have a discriminatory impact are not prohibited, and even if they are regarded as “high impact,” they are not disclosed to the public. In addition, CDT has made clear that the success of OMB’s guidance on agencies’ use and procurement of AI rests squarely on implementation – requiring effective resourcing, support, and transparency throughout the federal government.
Pursuant to the EO, individual agencies also released much-needed guidance to help companies institute common-sense governance measures that protect consumers from AI risks:
- The Department of Labor (DOL) released best practices for developers and employers to protect workers’ legal rights, establish AI governance (and human oversight), and ensure transparency and protect proper use of worker data when companies use AI. In addition, DOL released a guide for federal contractors and a bulletin on applying fair labor standards when using AI. CDT’s Workers Rights & Technology project has long called for these protections and underscored their importance in the modern workplace.
- The Department of Housing and Urban Development issued guidance making clear that housing providers have an obligation to protect renters from discrimination — both when advertising housing and making rental decisions. CDT has raised concerns about the role AI could play in perpetuating bias and discrimination in housing.
- The Department of Education published a toolkit for safe, equitable AI integration in K-12 public schools, detailing the long-standing civil rights laws that apply when districts acquire and use AI systems. CDT and other allies have documented the need for such guidance through an extensive body of work on the use of AI in education.
- The National Institute of Standards and Technology (NIST) formally launched the U.S. AI Safety Institute (AISI). CDT is a member of the AISI’s consortium, and has called for NIST to maintain a focus not only on long-term AI safety, but also measurement and evaluation for well-known areas of current risk.
- The Justice Department was tasked with leading coordination among enforcement agencies who address civil rights violations relating to AI, work CDT and allies have urged them to increase. The Department’s expected guidance on AI in the criminal justice system has not yet been publicly released.
In addition, the EO directed the National Telecommunications and Information Administration (NTIA) to evaluate the potential risks and benefits of open model weights. Consistent with CDT’s strong advocacy and coalition-building on the issue, NTIA issued its report in July urging the President not to place new restrictions on open model weights absent more evidence of risk, considering the clear benefits to competition, transparency, and security.
CDT also entered into a Memorandum of Understanding with the U.S. Access Board and the American Association of People with Disabilities to advance the interests of disabled people in the design and governance of AI.
“This one-year anniversary of the AI Executive Order isn’t the end of the road on AI governance—we’re still at the beginning,” said Givens. “The work ahead will involve moving beyond foundational guidance to implementing durable, effective AI governance, in both the private and public sectors. The Biden Administration should be praised for its cross-cutting approach to advancing responsible AI across the many different areas where AI is (or soon will) be used. Regardless of who is next in the White House, this work must continue to ensure America remains a leader in AI innovation that is responsible, trustworthy, and respects people’s rights.”
###