US AI Action Plan To Remove Misinformation, DEI References

MediaNama’s Take: The US aims to control AI technology and become a global leader in the AI world, as reflected in the sentiment of America’s AI Action Plan. While the US opposes AI models from “adversary nations”, claiming them as authoritarian and promoting censorship, Donald Trump follows a similar path by ordering the removal of references to misinformation, climate change, and DEI.

This directive contradicts existing responsible AI frameworks worldwide, which promote cultural representation and inclusion in AI systems. It can also increase friction among nations due to the lack of a unified ethical policy. To illustrate an example, if an Indian user asks US-based AI models about abortion, how would the models respond? The United States has tightened reproductive rights after the Supreme Court overturned Roe v. Wade. In this context, would the AI models lead the answers to reflect “American values,” or would they align with local culture and laws? How will AI companies ensure moral consistency for all users without bias?

While the US-based OpenAI’s ChatGPT has the most user base in India, global cooperation and coordination are very important to the existing businesses. The US has a strong, desperate desire to dominate the AI market. It is justified in seeking secure AI systems to defend itself and AI infrastructure against cyber exploitation by enemy nations. However, this need for secure AI systems and the urgency to be a global AI leader must be balanced. AI innovation should not be used as an easy excuse to bypass necessary regulations, including environmental clearances.

What’s the News?

US President Donald Trump has passed three executive orders and unveiled America’s AI Action Plan, a document with policy recommendations to help the nation achieve “global AI dominance.” The new policy comes after Trump replaced an executive order passed by President Biden in 2023.

This AI Action Plan comes at a moment when multiple American AI companies are partnering with the US government, especially with the defence and military departments. For example, OpenAI is leading the government efforts with initiatives like “OpenAI for Government.” Anthropic and Google also have deals with the Department of Defense to support AI systems for national security and intelligence. Similarly, Elon Musk’s xAI also partnered to offer generative AI for defense purposes, while Palantir, often through Accenture Federal Services, integrates AI-driven analytics into federal operations. Whereas, Scale AI aids in evaluating and adopting AI across agencies.

The AI Action Plan also focuses largely on using AI technologies in US defence and to strengthen AI infrastructure, including datacenters and manufacturing semiconductors. It offers strict export restrictions against non-US allied nations, emphasising on denying access to advanced technology to the adversary nations. At the same time, the policy also aims to ease a lot of regulations within the US that can hinder AI innovation.

In the name of encouraging AI innovation, Trump’s latest policy aims to encourage AI innovation by directing NIST to deprioritise misinformation in its AI Risk Management Framework. The policy also instructs federal agencies not to procure AI models with “ideological biases.” It orders to eliminate the references to Diversity, Equity, and Inclusion (DEI) and climate change, arguing that this is necessary to prevent social engineering agendas.

“As our global competitors race to exploit these technologies, it is a national security imperative for the United States to achieve and maintain unquestioned and unchallenged global technological dominance. To secure our future, we must harness the full power of American innovation.”
– Donald Trump in his Statement on AI Action Plan

What are the ‘AI Action Plan’ Executive Orders?

In the USA, the President can issue legally binding executive orders to exercise executive power by directing federal agencies and officers on how to implement laws and carry out their functions and duties.

As part of America’s AI Action Plan, Donald Trump has passed the following executive orders to achieve the President’s vision of “global AI dominance.”

Preventing Woke AI in the Federal Government

Trump signed this order to discourage the procurement of AI models which are ideologically biased towards “diversity, equity, and inclusion” (DEI). As AI will be used in learning new skills and to consume information, the executive order says that these can “distort the quality and accuracy of the output.”

“In the AI context, DEI includes the suppression or distortion of factual information about race or sex; manipulation of racial or sexual representation in model outputs; incorporation of concepts like critical race theory, transgenderism, unconscious bias, intersectionality, and systemic racism; and discrimination on the basis of race or sex,” reads the order. The order also claims that DEI-favoured AIs displace the commitment to truth in favour of preferred outcomes.

Accelerating Federal Permitting of DataCenter Infrastructure

Trump ordered the easing of federal regulatory burdens to accelerate the construction of AI datacenters. This order includes enabling the administration to use federally owned land for the fast and efficient development of datacenters.

Revoking the earlier executive order passed in January 2025, this order directs the Secretary of Commerce to consult with the Director of the Office of Science and Technology Policy (OSTP) to launch initiatives to provide financial support for Qualifying Projects.

The Secretaries of Defense, Interior, Commerce, or Energy can designate a company’s datacenter “Qualifying Project,” if it meets the following requirements:

The Project Sponsor has to commit at least $500 million in capital expenditures, as determined by the Secretary of Commerce.
Adds an electric load greater than 100 MW.
Protects national security.

Within 10 days of the order’s issuance, each relevant agency must identify to the Council on Environmental Quality any categorical exclusions established or adopted under the National Environmental Policy Act (NEPA). The order aims to facilitate the construction of Qualifying Projects through these exclusions.

Promoting The Export of the American AI Technology Stack

To decrease international dependence on AI technologies developed by their adversaries, Trump passed an executive order to ensure that American AI technologies and governance models are adopted by the US’s allies and to secure their technological dominance.

Within 90 days of the order’s issuance, the Secretary of Commerce is directed to consult the Secretary of State and the Director of the Office of Science and Technology Policy (OSTP) to establish and implement the American AI Exports Program to support the development and deployment of ‘full-stack AI export packages from the United States.’ This refers to complete, end-to-end solutions of AI technologies developed and packaged for export to other countries. Full-stack means that it includes all layers or components required to make the AI technology functional, from the infrastructure (like data centers) and hardware (like chips, servers, and accelerators) to the software and applications.

What does the AI Action Plan say?

The AI Action Plan also focuses largely on using AI technologies in US defence and to strengthen AI infrastructure, including data centers and manufacturing semiconductors. It offers strict export restrictions against non-US allied nations, emphasising on denying access to advanced technology to the adversary nations. At the same time, the policy also orders to ease a lot of regulations within the US that can hinder AI innovation.

In the name of encouraging AI innovation, Trump’s latest policy aims to encourage AI innovation by directing NIST to deprioritize misinformation in its AI Risk Management Framework. The policy also instructs federal agencies not to procure AI models with “ideological biases.” It orders to eliminate the references to Diversity, Equity, and Inclusion (DEI) and climate change, arguing that this is necessary to prevent social engineering agendas.

On Ideological Biases:

Ensuring AI Models Protect Free Speech while Deprioritising Misinformation

The AI Action Plan calls for AI models, particularly those procured by the US federal government, to preserve and uphold free speech. It also directs agencies to deprioritise and to remove references to “misinformation” from existing risk management frameworks.

The Department of Commerce (DOC) is asked to deprioritise the NIST AI Risk Management Framework to remove references to misinformation, Diversity, Equity, and Inclusion, and climate change.
Federal procurement guidelines will be updated to ensure the government contracts only with large language model (LLM) developers who maintain objectivity and avoid top-down ideological bias in their systems.

The DOC, through NIST’s Center for AI Standards and Innovation (CAISI), will research and publish evaluations of frontier AI models from the People’s Republic of China to assess alignment with Chinese Communist Party narratives and censorship

Encouraging Open-source and Open-weight AI Models based on ‘American Values’

While acknowledging that it is the right of the developer to release a certain model as closed-source or open-source, it suggests that the deferral government should create a supportive environment for open models.

Specific Recommendations:

Improving the financial markets to ensure that startups and academicians have access to large-scale computing power. It aims to do this by leveraging spot and forward markets, a standard trade practice where the payment can be agreed to be paid in the future after the delivery.
Establishing National Artificial Intelligence Research Resource (NAIRR) to connect nationwide researchers and educators nationwide with essential AI resources. It also asks to publish new AI breakthroughs by a National AI Research and Development (R&D) Strategic Plan.

On Reducing the Regulations:

Easing Regulations to Advance AI Innovation

The policy recommends removing bureaucratic red tape for the private sector. To avoid wasting funds, it also prohibits government funding for states with burdensome AI regulations. It argues that undue restrictions hinder innovation.

Specific Recommendations:

The Office of Science and Technology Policy (OSTP) is a recommended request for information from businesses and the public on Federal regulations that are hindering AI innovation and adoption, and it also asks to collaborate with relevant agencies to take necessary action.
Similarly, the Office of Management and Budget (OMB) is to comply with the earlier executive order “Unleashing Prosperity Through Deregulation,” to work with all Federal agencies and identify, revise, or repeal regulations, rules that unnecessarily make AI development harder.
The order also directed the Office of Management and Budget (OMB), Federal Communications Commission (FCC), and Federal Trade Commission (FTC) to assess and investigate measures that could prevent AI innovation with undue burdens.

On AI Adoption:

Encouraging the Government to use AI to accelerate manual internal processes

Formalising the Chief Artificial Intelligence Officer Council (CAIOC) as the main platform for inter-agency coordination and collaboration in AI adoption.
Creating a talent-exchange program within government workspaces

Enabling AI Adoption with “try-first” culture.

The policy recommends a “try-first” culture for AI across American industries to drive faster adoption, even in critical sectors like healthcare. It cites distrust, a lack of understanding of the technology, a complex regulatory landscape, and unclear governance and risk mitigation standards as reasons for slow AI adoption.

Specific Recommendations:

Establish regulatory sandboxes or AI Centers of Excellence nationwide to allow researchers, startups, and enterprises to test and adapt AI tools for the Department of Defense (DOD) and Intelligence Community (IC) AI adoption initiatives.
Assessing AI adoption in national security sectors and their security implications to compare the US with its adversaries.

Upskilling American Workers with AI Literacy

Directed by the Department of the Treasury to offer tax-free reimbursement for AI-related training.
Monitoring and studying the labour market for the impact of AI in the workforce and business.
Directed the Department of Labour to use discretionary funding to provide retraining for individuals impacted by AI-related job displacement.

The AI Action Plan On Investments by the US

Encouraging and Investing in the AI-led Manufacturing Sector

It says that AI will enable the new technology across industries, including autonomous drones, self-driving cards and even those technologies which don’t have terminology yet. So, it suggested that the government should invest in the emerging AI-led “industrial renaissance.”

Invest in AI-Driven Scientific Research and Infrastructure

Investment in automated cloud-enabled labs in scientific fields, including engineering, materials science, chemistry, biology, and neuroscience.
Incentivise researchers to publicly release high-quality datasets.
Require federally funded researchers to disclose non-proprietary, non-sensitive datasets used in AI models during research and experimentation

The AI Action Plan In Defence and Military

Encouraging the usage of AI in the Department of Defence

“AI has the potential to transform both the warfighting and back-office operations of the DOD,” reads the policy and recommends aggressive AI adoption that is secure and reliable.

Identifying the talent and skills needed for the Department of Defense (DOD) workforce to leverage AI systems by implementing talent development programs.

Transform Senior Military Colleges into hubs for AI research, development, and talent building, integrating AI-specific curriculum across all majors.

Prioritising Department of Defense (DOD) agreements with cloud service providers, computing infrastructure operators, and other private sector entities to secure priority access to resources during a national emergency.

Invest in increasing the predictability of AI to use it in defence and national security.

They cite the lack of an accurate behaviour prediction system of an AI model as the challenge preventing the U.S. from using advanced AI in high-stakes defense and national security.

Specific Recommendations:

AI hackathon initiative with federal agencies and academicians to test AI systems for transparency, effectiveness, use control, and security vulnerabilities.

Including AI interpretability, control, and robustness as key priorities in the forthcoming National AI Research and Development (R&D) Strategic Plan.

Building High-Security DataCenters for Military and Intelligence

“Because AI systems are particularly well-suited to processing raw intelligence data today, … it is likely that AI will be used with some of the U.S. government’s most sensitive data,” states the policy. It also mandated that the datacenters where the AI models are deployed must resist attacks by adversary nation-state actors.

Specific Recommendations:

Establishing the new technical standards for high-security AI datacenters and promoting the adoption of classified compute environments to ensure secure and scalable AI operations.

Adopting AI Systems as cyberdefence tools

Encouraging the AI adaptation in governmental operations, the paper reads, “Fortunately, AI systems themselves can be excellent defensive tools.”

Establishing an Artificial Intelligence Information Sharing and Analysis Center (AI-ISAC) to promote the sharing of AI security threat information among the US’s sectors with critical infrastructures
The Department of Homeland Security (DHS) is directed to issue and maintain guidance for the private sector addressing AI vulnerabilities and threats. The federal agencies are also directed to collaborate to share AI vulnerabilities with the private sector as well.
Promoting Secure-By-Design AI Technologies to avoid data poisoning and privacy attacks.

It says that the US government has the responsibility to ensure that AI systems that the state relies on have to be protected against spurious or malicious inputs.

Specific Recommendations:

Directs the National Institute of Standards and Technology (NIST) and the Director of National Intelligence (ODNI) to continue refining DOD’s Responsible AI and Generative AI Frameworks, roadmaps, and toolkits.
It also directs the Cybersecurity and Infrastructure Security Agency (CAISI) to publish an Intelligence Community (IC) Standard on AI Assurance under Intelligence Community Directive 505 on Artificial Intelligence.

Integrating AI Incident Response into National Security Protocols

US to lead efforts in evaluating national security risks in Advanced AI models.

“The most powerful AI systems may pose novel national security risks in the near future in areas such as cyberattacks and the development of chemical, biological, radiological, nuclear, or explosive (CBRNE) weapons, as well as novel security vulnerabilities,” reads the policy. The policy emphasises the need to understand these risks as they develop to protect national defense and homeland security.

The policy recommends prioritisation of the recruitment of top AI researchers within federal agencies.
It orders to evaluate frontier artificial intelligence systems for national security risks by collaborating with the Center for Artificial Intelligence and Security Innovation (CAISI) along with other agencies that have expertise in chemical, biological, radiological, nuclear, and explosive (CBRNE) threats and cyber risks.
The policy orders to assess potential security vulnerabilities and malign foreign influence through adversary AI systems in critical infrastructure and the US economy.
Investment in biosecurity to avoid synthetic pathogens and biomolecules

The policy recommends establishing systems to screen biological sequence orders before fulfilling them, after approving verified customers, to detect and prevent fraudulent or malicious activity.

Specific Recommendations:

It orders that the Office of Science and Technology Policy (OSTP) should lead efforts to convene government and industry stakeholders to develop data-sharing systems among nucleic acid synthesis providers to detect fraudulent or malicious customers.
It requires institutions receiving federal funding for scientific research to use nucleic acid synthesis tools and providers with robust sequence screening and customer verification procedures.

The AI Action Plan On AI Reliability

Evaluations to assess the performance and reliability of AI systems.

Publishing guidelines and resources to help federal agencies evaluate AI systems for mission-specific operations and legal compliance.
Investment in developing AI testbeds to pilot AI systems in secure, real-world settings. These testbeds should allow researchers to prototype new AI systems and make them market-ready.
An AI Consortium to establish new measurement science for identifying proven, scalable, and interoperable techniques and metrics to advance AI development.

The AI Action Plan for Semiconductors and Infrastructure

Restoring American Semiconductor Manufacturing Business

The policy recommends bringing semiconductor manufacturing back to US soil to reinforce technological leadership, to control the supply chain and disruption from foreign rivals.

It recommends the Department of Commerce’s (DOC) revamped CHIPS Program to focus on strong returns on American taxpayers’ investment, and it also recommends eliminating unnecessary policy requirements for CHIPS-funded semiconductor manufacturing projects.
The Department and relevant federal agencies should collaborate to streamline regulations that hinder semiconductor manufacturing.
Permitting datacenters and semiconductor manufacturing facilities while maintaining a steady energy supply.

“America’s environmental permitting system and other regulations make it almost impossible to build this infrastructure in the United States with the speed that is required,” states the paper.

The document states that Trump’s revised reforms of the National Environmental Policy Act (NEPA) ease the regulations across the relevant agencies.

Specific Recommendations:

Establishing new Categorical Exclusions under the National Environmental Policy Act (NEPA) for datacenter actions with minimal environmental impact.
Expand the FAST-41 process to cover all eligible datacenter and energy projects. It sets deadlines and standardised steps to reduce delays and redundancies. It helps critical projects secure permits and get built more quickly.
Similarly, exploring the need for a nationwide Clean Water Act’s Section 404 permit for datacenters. This also ensures that it simplifies and accelerates the authorisation process.
Making federal lands available for the construction of datacenters and power generation infrastructure by directing agencies to identify suitable large-scale development sites.

The AI Action Plan On Diplomacy

Exporting AI-stack to American Allies

The AI technology stack includes hardware, models, software, applications, and standards for all American allies and for those countries which are willing to join its alliance. The policy also states that refusing to join the alliance would turn the countries into US rivals.

Specific Recommendations:

The Department of Commerce (DOC) is ordered to establish a program to collect proposals from the industry for full-stack artificial intelligence (AI) export packages. The departments are also directed to ensure they facilitate deals that meet US-approved security standards and requirements.

Countering Chinese Influence in International Governance Bodies

The US says it supports the development of artificial intelligence in like-minded nations that share American values. However, it argues that other nations’ burdensome regulations and vague “codes of conduct” promote cultural agendas that conflict with American values, attempting to alter standards for facial recognition and surveillance.

Specific Recommendations:

It directed the Department of State (DOS) and the Department of Commerce (DOC) to use the US’s position in international diplomacy to advocate for AI governance approaches that promote innovation, reflect American values, and counter authoritarian influence.

Denying AI Computing Power to Adversary States

“Denying our foreign adversaries access to this resource, then, is a matter of both geostrategic competition and national security,” reads the policy.

Specific Recommendations:

The AI Action plan directed to utilise existing location verification features to ensure chips used on advanced artificial intelligence (AI) computers do not end up in “countries of concern.”
Monitoring emerging AI compute technologies to identify potential diversion risks in various countries or regions. It also ordered that enhanced monitoring should be implemented to check end-use cases in countries with a high risk of diversion, particularly where the Bureau of Industry and Security Export Control Officer is not present.

The AI Action Plan On Exports

Controlling the Exports to prevent Adversary nations from using the US Tech

Emphasising on dominating manufacturing and research, the policy states that the US must prevent adversaries from using US innovations for their benefit, as it could undermine the US’s national security. It also states that the United States and its allies currently restrict exports of major systems critical to semiconductor manufacturing. But, lacks control in many of the semiconductor’s sub-component systems.

The policy directs the Department of Commerce (DOC) to develop export controls on semiconductor manufacturing systems,

Strong export controls on sensitive technologies

The policy suggests using the Foreign Direct Product Rule and secondary tariffs to control the exports of technology. This is to ensure that U.S. adversaries do not receive advanced technologies.

Specific Recommendations:

It ordered the Department of Defence to develop new controls to prevent US adversaries from accessing the U.S. defence-industrial base and acquiring stakes in companies which deal with defense suppliers.
It instructed its departments to develop, implement, and share complementary technology protection measures, especially in basic research and higher education, to mitigate risks from strategic adversaries.
It suggests that this plan should align the incentives and policy to encourage allies to adopt AI protection systems to control the controls throughout the supply chain.
It said that the US should expand its initiatives for plurilateral controls on the AI technology stack to avoid the reliance on multilateral treaty bodies.

Also Read

Support our journalism: