AI Week in Review - 4/26/25

Public Sector

Author

Chris Kraft
April 26, 2025

The Highlights

The Story
The Executive Order aims to ensure the United States remains a global leader in AI by:

  • Promoting AI literacy and early exposure for students.

  • Training educators to teach and use AI effectively.

  • Establishing public-private partnerships to support K-12 AI education.

  • Expanding AI-related apprenticeships and workforce training programs.

  • Launching a Presidential AI Challenge to showcase student and educator achievements.

More Details
This will be accomplished by doing the following:

  • Creating an AI Education Task Force

  • Launching the Presidential AI Challenge

  • Improving AI Resources for K-12 Education

  • Enhancing Educator Training in AI

  • Expanding AI-Related Apprenticeships

  • Broadening Federal Fellowship and Scholarship Support

The Story
The White House Office of Science and Technology Policy (OSTP) just released public comments (over 10,000) in response to the RFI on the development of the AI Action Plan. These comments come from a diverse group of individuals, companies, civil society, and others. The consolidated PDF is over 18,000 pages.

The Details

Federal

Trump signs Executive Order to ramp up K-12 AI education
President Trump’s new Executive Order aims to expand K-12 AI education nationwide, but education leaders stress the need for sustainable funding and greater educator involvement in shaping its rollout.

American public submits over 10,000 comments on White House’s AI Action Plan
In response to President Trump’s AI executive order, the White House released over 10,000 public comments that will help shape America’s AI Action Plan, reflecting broad national engagement on key AI priorities.

Air Force releases new doctrine note on AI to guide future warfighting
The U.S. Air Force released Doctrine Note 25-1 to guide Airmen in ethically and effectively integrating AI into operations, from combat to logistics, amid growing national defense priorities.

The government embraces AI lab rats
The FDA is beginning to replace animal testing with AI and digital models, aiming to reduce costs, accelerate drug development, and limit animal use in preclinical research.

GenAI's environmental and human effects
GAO’s latest report warns that while generative AI can boost productivity, it also poses unclear but significant environmental and societal risks—prompting policy options to improve oversight, data transparency, and ethical use.

Defense officials outline AI's strategic role in national security
U.S. Defense leaders emphasize AI’s growing role in military readiness, cybersecurity, and operational decision-making, aligning with a broader push to modernize national security through accelerated, ethical AI adoption.

IARPA looks to next round of AI cybersecurity research
As IARPA concludes its TrojAI program, it’s shifting focus to securing generative AI models for intelligence use, aiming to prevent classified data leaks from systems trained on sensitive information.

The top three priorities for federal CIOs in 2025: Navigating budget cuts with GenAI
Facing budget cuts and workforce challenges, federal CIOs are urged to harness GenAI to streamline operations, enhance citizen services, and optimize unstructured data for greater efficiency and resilience.

Veteran suicide prevention increasingly looks to use AI
The RAND Corporation found that veteran suicide prevention programs are increasingly adopting AI and digital tools, with 37% of proposed initiatives planning to use AI for risk identification and real-time monitoring

State / Local

NJ advances AI through an economic development
New Jersey unveils two AI initiatives—offering grants and tax incentives—to attract private-sector investment, boost job creation, and establish the state as an East Coast AI innovation hub.

Agentic AI gives Oklahoma cyber ops a powerful but ‘scary’ tool
Oklahoma has deployed autonomous AI agents in its cybersecurity operations to counter escalating AI-driven threats, balancing speed and scale with risks of reduced human oversight.

PA lawmakers look to set guidelines on safe AI development
Pennsylvania legislators are launching a statewide initiative to shape AI policy through hearings, committee work, and proposed laws aimed at balancing innovation with safeguards against misuse.

State and local agencies deploy AI for document processing
State and local governments are using AI-powered document processing to automate data extraction, speed up services, and free staff for higher-value work, while maintaining human oversight.

How good is state AI guidance? Depends where you live
Most states' AI guidance for schools overlooks key concerns like deepfakes and community engagement, with only a few encouraging meaningful parental involvement in policymaking.

Agentic AI: The future of government operations and citizen services
Agentic AI can transform public service delivery by automating routine tasks, accelerating decision-making, and enabling civil servants to focus on human-centric roles that improve citizen outcomes.

California Supreme Court demands State Bar answer questions on AI exam controversy
The California Supreme Court is demanding answers from the State Bar after it quietly used AI to help create bar exam questions, raising concerns over transparency, validity, and oversight in high-stakes testing.

Police apprehensive about AI as surveillance tech grows
A University of Michigan survey found rising use of surveillance tech among Michigan law enforcement, but widespread uncertainty and skepticism about adopting AI tools for criminal justice tasks.

North Dakota law expected to curb AI influence in health care
North Dakota’s new law restricts AI involvement in health care prior authorizations, requiring licensed physicians to make final decisions to reduce treatment delays and prioritize patient care.

Univ of Michigan, Google launch agentic AI for Tas
The University of Michigan and Google Public Sector launched an AI-powered virtual teaching assistant pilot to enhance student engagement and inform a 26-campus study on agentic AI in education.

Pittsburgh electricity supplier warns state on AI energy needs
A Duquesne Light executive warned Pennsylvania regulators that unchecked AI-driven data center energy demands could strain Pittsburgh’s grid and shift infrastructure costs onto regular customers.

How the state is turning to AI to help detect Medicaid fraud
Louisiana is deploying AI technology developed with the University of Louisiana at Lafayette to detect Medicaid fraud, aiming to improve program integrity alongside new maternal overdose prevention initiatives.

International

UAE wants to be the first in the world to write laws using AI
The UAE is launching the world’s first initiative to use AI not just to support, but to actively rewrite and update laws—aiming to boost legislative speed and precision by 70%.

AI boosts diplomacy—but escalates cyber risks, warns UK Foreign Office
The UK Foreign Office acknowledges AI's power to boost diplomatic efficiency while warning it will also amplify cyber risks, prompting evolving security policies and risk management efforts.

Tackling legal risks in GenAI
Singapore’s IPOS International urges public sector leaders to address GenAI’s legal risks—from copyright and data protection to liability and contracts—to ensure responsible innovation.

How Dubai is using AI to reimagine public services
Dubai is pioneering AI integration across 15 key government functions—from customer service to heritage preservation—demonstrating how AI can enhance efficiency, inclusion, and innovation in public service delivery.

Brazil’s AI-powered social security app is wrongly rejecting claims
Brazil’s AI-driven social security app is wrongly denying benefits to vulnerable citizens due to errors and lack of human oversight, disproportionately impacting rural and digitally excluded populations.

India’s AI-driven healthcare transformation
India is transforming healthcare by leveraging AI-driven drug discovery, digital tools for patient engagement, strategic investment, and interoperable health data to create scalable, inclusive innovations for emerging economies.

Lessons from Japan: Governing in the age of AI
Japan’s human-centric, transparent, and strategically directed AI approach offers crucial lessons for countries like the UK on modernizing public services while maintaining public trust and accountability.

Malaysia aims to be a global leader in responsible AI innovation
Malaysia’s National AI Office is advancing a dual strategy of accelerating AI adoption across industries while strengthening ethical governance to position the country as a global leader in responsible AI innovation.

Estonia moves to develop new cross-government, AI-powered data management asset
Estonia is launching a cross-government, AI-powered data management platform to enhance citizen access to public services, building on its long-standing commitment to digital innovation and security.

Peru's AI regulatory boom: Quantity without depth?
Despite a surge in AI-related legislation, Peru’s regulatory efforts lack enforceable protections and institutional depth, risking superficial governance and potential harms in a fragile democratic context.

How to move the needle on AI adoption
Singapore’s new Enterprise Compute Initiative (ECI) aims to accelerate AI adoption across enterprises by offsetting development costs, providing expertise, and helping businesses move beyond pilot projects toward industry-wide transformation.

A deep peek into DeepSeek AI’s talent
DeepSeek AI’s rise reveals China’s growing ability to develop world-class AI talent domestically, challenging longstanding U.S. assumptions about its dominance in global AI innovation.

Everything Else

The AI gap in executive leadership teams
Executive teams with strong AI literacy outperform peers by aligning strategy with AI capabilities, avoiding hype, and making smarter decisions through shared understanding and collective learning.

What separates AI leaders from laggards
Companies moving past AI pilot paralysis are embracing blended teams of full-time and freelance talent, building custom solutions with open-source tools to accelerate innovation and realize ROI.

Does it cost more money to be polite to AI chatbots?
Politeness to AI chatbots slightly increases energy use, with OpenAI confirming that even small phrases like "please" and "thank you" add up to significant processing costs and environmental impact.

CIOs increasingly dump in-house POCs for commercial AI
Facing high failure rates and resource constraints, CIOs are abandoning homegrown AI pilots in favor of commercial tools as vendors rapidly embed generative AI into enterprise software.

AI outsmarts virus experts in the lab
A new study reveals advanced AI models now outperform virologists in lab problem-solving, prompting urgent calls for safeguards to prevent misuse in bioweapon development.

Microsoft thinks AI colleagues are coming soon
Microsoft predicts 2025 will mark a shift toward “Frontier Firms” using AI agents as digital colleagues, transforming operations and accelerating business innovation across industries.

Anthropic considers model welfare
Anthropic has launched a research program to explore the ethical and scientific challenges of AI model welfare, acknowledging the growing need to consider the experiences and moral status of advanced AI systems.

BOA’s big bet on AI started small
Bank of America’s $4 billion AI investment builds on years of practical, small-scale innovation like its Erica virtual assistant, prioritizing trusted, scalable AI applications over chasing flashy new technologies.