NewDEAL Forum AI Task Force Letter to White House Office of Science and Technology Policy
In response to a February 2025 request for public comment from the White House Office of Science and Technology Policy on the Trump Administration’s AI Action Plan, NewDEAL Forum AI Task Force co-chairs Dorcey Applyrs, Alex Bores, and Matt Mahan submitted the following letter.
Dear members of the White House Office of Science and Technology Policy,
We are responding to the Request for Information regarding recommendations and input for the Development of an Artificial Intelligence (AI) Action Plan, as outlined and directed by the January 23, 2025, Presidential Executive Order.
It is our desire that the final action plan accomplish one overarching goal: To harness the tremendous benefits of AI for both the public and private sectors while mitigating potential harms.
We come to you as both elected officials in our respective cities/states as well as co-chairs of the NewDEAL Forum AI Task Force.
The NewDEAL Forum is a nonpartisan 501(c)3 organization which identifies and elevates innovative, future-oriented state and local policies that can improve the lives of all Americans. By facilitating the identification and spread of policy ideas, the NewDEAL Forum seeks to foster economic growth, reduce barriers to opportunity, and promote good government in communities, cities, and states throughout the country.
The AI Task Force launched in the summer of 2024, and has since brought together state and local elected officials along with experts from non-profits and the private sector to explore AI's potential to make government work better. The group’s efforts underscore the significant role and opportunities for state and local governments when it comes to advancing the potential of AI while mitigating potential harms.
In 2024, the Task Force released a report containing recommendations on Election Security (Artificial Intelligence and Elections: Initial Steps To Protect Democracy).
While the Task Force is largely focused on state and local initiatives related to AI, we recognize the need for the federal government to provide overarching guidance on this emerging technology. We believe that federal agencies can and should set the example of AI innovation, showing how the technology can make government more efficient.
We are encouraged by efforts that are already underway. For example, Arati Prabhakar, former Director of the White House Office of Science and Technology Policy, spoke in December about how the Social Security Administration is using AI to speed up the application process. In addition, the Department of Veterans Affairs is using AI to help collate and digest feedback from veterans around the country. We hope the current administration builds on these efforts.
The recommendations in this letter are drawn from the experiences of state and local elected officials across the nation, as well as our conversations with private sector and non-profit experts in the field. It is our sincere hope that this administration provides a policy framework that both helps secure a brighter future for American families while also providing appropriate guardrails to protect these same families from potential harm, all while allowing AI technology to flourish.
Our recommendations are organized into three areas of focus: Making Government More Efficient; Election Security; and Regulation.
Innovation and Making Government More Efficient: Federal agencies must set the example of responsible AI integration. As noted above, federal agencies like the Social Security Administration and Department of Veterans Affairs are already using AI to increase efficiency. Federal agencies should continue to innovate with AI with particular emphasis on data security and privacy.
Focus on data privacy and security. The federal government should set the gold standard when it comes to data privacy and security. The potential of AI to create efficiencies and enable technological breakthroughs is tremendous, but guardrails are critical to protecting the privacy of individuals. For example, integrating AI technology into health care fields could prove beneficial, but personally identifiable information (PII) and health care information about individuals must be protected. OSTP must ensure that any federal agency using AI has clear, consistent, and strong protections for data privacy, including but not limited to preventing sensitive, HIPAA (Health Insurance Portability and Accountability Act) protected information from being disclosed.
Promote Collaboration and Information Sharing: OSTP should collect and disseminate best practices not only for other federal agencies, but also for state and local governments, industry stakeholders, and civil society organizations. The federal government should make a dedicated, concerted effort to work with state and local governments specifically to foster a shared understanding of AI challenges and opportunities when it comes to government efficiency. In addition, the federal government should learn from states and localities who are implementing AI effectively. Specifically, OSTP should take a leading role in convening stakeholders from state and local governments, non-profit groups, the private sector, and federal agencies to learn from each other and regularly share best practices. In addition, OSTP should consider other means of accomplishing this goal, including working groups, task forces, and white papers. The broad goal should be for high-quality AI use while maintaining data privacy and security. The reciprocal learning and sharing will benefit everyone involved and ultimately benefit the American people.
Incentivize AI Adoption: In our conversations with experts, we understand that many Americans are hesitant to adopt AI technology in part because it is an unknown technology to many. OSTP, along with other entities in the federal government, can and should incentivize state and local governments to provide assistance and opportunities for employees to understand and implement AI initiatives related to creating more government efficiencies and impactful services. The more the American people see the benefits of AI in conjunction with strong data privacy guardrails, the more comfortable they will be. States are said to be the laboratories of democracy, and emerging technologies like AI are a great platform for states and localities to experiment with various uses.
AI and Election Security: Election security is of paramount importance. Americans deserve to know that the political ads they see and hear are genuine rather than deepfakes created to confuse and misinform. Further, Americans need to continue to trust that our elections are safe from cyberattacks, including AI-enhanced cyberattacks. To safeguard the integrity of our electoral processes, the federal government must take decisive action to mitigate risks posed by AI while not impinging on First Amendment protections of Americans. Here are four concrete actions the federal government, with OSTP’s leadership, can implement:
Mandate Disclaimers for AI-Generated Political Content: The federal government should establish a nationwide standard requiring clear and conspicuous disclaimers on all AI-generated content used in political campaigns. This includes, but is not limited to, deepfakes, synthetic media, and AI-generated audio or video impersonations. This can be accomplished in ways that do not hamper the ability of campaigns to use AI but prevent the malicious and dishonest use of technology to confuse and misinform voters.
Support Rapid Response Mechanisms: Elections are largely run at the state and local level. Over the decades, we expect more and more from state and local election officials, but ensuring strong and secure elections takes adequate resources. The federal government can and should help by providing funding and resources to state and local election officials to establish rapid response teams capable of quickly identifying and countering AI-generated disinformation campaigns. These response teams would be non-partisan, with the goal of preventing the spread of AI-generated misinformation before it could garner a foothold and influence voters. Fair and secure elections are in our national interest.
Establish a Content Verification System: Voters deserve to know the truth, especially when it comes to voting for representatives, whether on city council or in Congress. Yet AI has the potential to confuse and mislead voters with deceiving content that looks real but contains knowingly false information. To combat this, the federal government, led by OSTP, should invest in the adaptation of a "digital marker" or similar verification system to ensure the authenticity of official election-related communications. One promising approach is the C2PA standard, an open-source technical specification endorsed by leading tech companies and some states. C2PA embeds provenance data directly into media content, allowing anyone to verify its origins and track any edits or manipulations. This “marker” could be used on media posts sharing information about the dates and hours when polling locations are open, thus giving American citizens the confidence to know the information is valid. The more confidence voters have in our elections, the stronger our democracy will be.
Increase Funding for Cybersecurity: For all the potential benefits of AI technology, there are malign actors who will attempt to use it for nefarious and illegal purposes, especially around elections. As noted, before, states and cities are largely responsible for administering elections but need adequate funds to do so. OSTP and the federal government must show that they care as much about mitigating potential harms of AI as they do about promoting the efficiencies the technology can unleash. Therefore, the federal government, with OSTP’s support, should allocate additional funding to enhance cybersecurity measures for state and local election infrastructure to protect against AI-driven cyber threats.
AI Governance and Regulation: It is possible to support and advance AI technology while still maintaining adequate guardrails to mitigate potential harm. For example, seat belt and air bag technologies improved automobile safety without compromising innovation and advancement. In the same way, the federal government can support and promote AI growth while taking necessary steps to mitigate potential harms through appropriate regulations.
Implement Algorithmic Accountability Measures: AI, at its base, is reliant on data. And the output from AI is only as good as the data and algorithms that go into it. As with any emerging technology, there will be glitches along the way, and the federal government can put in guardrails to reduce the harm of potential glitches. As such, OSTP should require companies and government agencies to conduct regular audits of their AI systems to ensure they are free from bias. This includes, but is not limited to, both political and religious bias, so that AI technology does not discriminate against anyone. This will make Americans of all faiths, creeds, nationalities, races, and political affiliations more readily embrace AI technology, which is necessary to ensure Americans are leading the way on AI.
Prioritize Equity and Inclusion: In the same way that the creation and adoption of email changed how all Americans interact with each other, AI technology has the potential to be just as transformative. If America is going to be dominant in the global AI landscape, we must ensure that the technology addresses the needs of ALL Americans. This means that AI policies must be developed and implemented in a way that promotes equity and inclusion, addressing the needs of traditionally underserved populations. The more people and communities who embrace the benefits of AI, the more leadership America has on the global stage. It would be a mistake to believe that America could lead the way on AI technology while leaving some communities behind.
This document is approved for public dissemination. The document contains no business-proprietary or confidential information. Document contents may be reused by the government in developing the AI Action Plan and associated documents without attribution.
Thank you for your time and consideration of these recommendations.
Sincerely,
Dr. Dorcey Applyrs
Chief City Auditor, Albany, NY
Alex Bores
District 73 Assemblymember, New York
Matt Mahan
Mayor, San Jose, CA
