AI Week in Review - 3/22/25

Public Sector

Author

Chris Kraft
March 22, 2025

Here’s your latest public sector AI news. There have been a lot of changes in the public sector recently - but the opportunities to leverage AI to support mission outcomes hasn’t changed. This week we see AI being used in healthcare, education, policy, and many other areas. We also have new technical capabilities from Anthropic - giving Claude the ability to perform web searches.

In this week’s edition:

  • State AI Roadmaps

  • PA Teachers Use AI

  • AI in the ER

  • Future of AI

  • Public (and private) sector AI news

The Highlights

Gemini Imagen3

The Story
State CIOs are being urged to build AI strategies that prioritize responsible governance, data quality, and workforce training. With generative AI gaining momentum, best practices from national associations offer guidance to reduce risks and maximize impact.

More Details

  • Governance matters: NASCIO recommends AI roadmaps include clear oversight, strong data sourcing, and workforce training.

  • Accountability first: AAAS suggests states adopt risk management, publish AI use cases, and create transparent acquisition policies to promote trustworthy AI.

  • Data quality is key: NGA and AAAS warn that hallucinations, bias, and weak safeguards in generative AI require strong data governance to ensure safe, accurate outcomes.

  • Workforce readiness: Pennsylvania piloted ChatGPT Enterprise; New Jersey launched statewide AI training. Surveys show public employees are eager to learn AI skills, even if most haven’t used the tools yet.

  • Bottom line: States must align strategy, safeguards, and skills to responsibly scale AI in government.

The Story
Pennsylvania teachers are embracing AI to boost student engagement, streamline lesson planning, and prepare students for a tech-driven future — while balancing concerns about misuse and equity.

More Details

  • Richland High uses Midjourney to bring history lessons to life and supports a student gaming club with AI-generated visuals.

  • Teachers use AI to check curriculum gaps, provide student writing feedback, and create immersive learning experiences.

  • An “Exploring AI” class teaches students how to use, train, and critically evaluate AI tools, with an emphasis on ethics.

  • Schools are adopting AI policies focused on responsible use, privacy, and safety — following both excitement and incidents of misuse.

  • Educators see AI not as a replacement but as a tool to enhance learning, saying the future belongs to those who can use it wisely.

The Story
Long emergency room wait lines, delayed diagnosis, and missed opportunities to streamline the overall process leads to unnecessary deaths. This report looks at the UK National Health Service (NHS), but it could be just as applicable in the United States or across the globe.

More Details

  • Long waits cost lives. In 2024, the Royal College of Emergency Medicine estimated that about 300 deaths per week were associated with long waits in the ER.

  • The AI-enhanced process includes seven main tasks - collecting information, filtering requests, assessing health conditions, prioritizing for treatment, identifying care services and next steps, allocating resources, and improving the health system.

  • Using AI throughout the process would also enable more dynamic navigation.

The Story
This report is packed with 23 leading voices in AI. While these perspectives can help startups, they are also great for those of us in the public sector. Dive into the most critical AI and Cloud AI innovations, trends, opportunities, and challenges that leaders should consider in 2025.

More Details
Here are my favorite observations.

  • Compute costs – Bet on compute being faster, cheaper, and more reliable.

  • Human productivity – Individuals will be up to 10x more productive with AI.

  • Don’t obsess over ROI – Use AI to unearth savings through efficiency.

  • Still early days in AI – despite all the hype, we are still in AI early days. The biggest opportunities may be in plain sight.

  • Don’t delay AI adoption – Delaying puts companies at risk of becoming irrelevant.

  • Go native – Become AI-native across all functions.

  • Don’t wait – You can’t afford to wait for model improvements before using them.

The Details

Federal

AI to boost efficiency, optimize logistics support as DLA
Defense Logistics Agency is scaling AI adoption with over 55 models in use, streamlining operations, reducing costs, and ensuring secure, data-driven decision-making across logistics and finance.

Top admiral weighs in on AI use in the Navy
Adm. Daryl Caudle highlights AI’s growing role in the Navy, aiding in personnel assessments, data analysis, and ship readiness to improve efficiency and operational effectiveness.

Inside industry’s wishlist for a new national AI strategy
Industry leaders urge a new U.S. AI strategy to enhance global dominance, streamline regulations, boost federal adoption, and balance innovation with competition, security, and international AI diplomacy.

Strategic vision for US AI leadership: Security, innovation, democracy and global prosperity
​The Wahba Institute for Strategic Competition emphasizes the need for the U.S. to maintain human oversight over AI, invest in talent and infrastructure, and collaborate with allies to ensure AI promotes democracy and global prosperity.

Space Force publishes 2025 Data, AI strategic action plan
The U.S. Space Force released its FY 2025 Data and AI Strategic Action Plan, outlining efforts to enhance digital fluency, AI literacy, and data-driven decision-making to maintain space superiority.

GSA debuts new generative AI tool for workers
The GSA has launched a generative AI chatbot to assist staff with routine tasks, aiming to foster AI adoption across government amid concerns over workforce reductions and surveillance.

USPS plans to use AI to enhance customer service
USPS plans to adopt generative AI and a cloud-based call center platform this summer to improve customer service and operational efficiency, starting with automating responses to common delivery inquiries.

To win the AI race, the US needs an all-of-the-above energy strategy
To maintain AI leadership over China, the U.S. must adopt an all-of-the-above energy strategy that rapidly expands power generation, modernizes the grid, and avoids limiting any energy source.

Government AI hire, use, buy
Georgetown’s Government AI HUB Roundtable series concluded with a call for centralized transparency, robust AI workforce investment, and bipartisan momentum to responsibly scale AI in U.S. government.

State / Local

Pomona, Calif., uses AI, cameras to address traffic safety
This AI-driven transportation project will use video cameras and smart streetlights to improve traffic flow, reduce accidents, and enhance emergency response along key roadways.

Pennsylvania educators find positive applications for AI
Pennsylvania educators are embracing AI for lesson planning, student engagement, and writing support while balancing responsible use, policy regulation, and ethics to prepare students for future careers.

Joint California Policy Working Group on AI frontier models
A report commissioned by Governor Newsom provides a policy framework for governing frontier AI in California, balancing innovation with risk mitigation through interdisciplinary research and strategic regulation.

State AI roadmaps must mitigate risks and build competent workforces
State CIOs are urged to prioritize AI governance, data quality, and workforce training in AI roadmaps to manage risks and build competent, future-ready public sector teams.

Illinois lawmakers considering several measures related to AI
Illinois lawmakers are considering multiple AI-related bills, addressing education, healthcare, elections, and worker protections, as they seek to balance regulation with innovation in a rapidly evolving landscape.

New York State names Chief AI Officer
New York State appointed Shreya Amin as its new Chief AI Officer to guide ethical AI adoption, enhance services, and strengthen oversight, reinforcing the state’s growing leadership in public sector AI innovation.

PA Governor says ChatGPT saves staff 8 hrs/week
Pennsylvania’s ChatGPT pilot saved employees eight hours weekly, proving AI can boost efficiency without replacing human expertise, according to Governor Josh Shapiro.

California’s United Way using AI to cut costs, improve services
Orange County’s United Way is using AI and predictive analytics to cut costs and proactively deliver social services, identifying early signs of homelessness and improving care coordination across 1,000+ providers.

Boston City Council to explore how AI could improve service
Boston City Council has approved a hearing to explore how AI can responsibly enhance city services, focusing on ethics, workforce impact, and real-world municipal applications.

California ‘No Robo Bosses Act’ would bar AI from making personnel decisions
California’s proposed “No Robo Bosses Act” would ban employers from using AI alone to make job decisions, mandating human oversight to prevent bias and protect worker rights.

NY goes old school – straps phones to metro cars
NYC’s MTA is partnering with Google Public Sector to test AI-powered subway track monitoring using smartphones, aiming to reduce delays while keeping human inspectors in the loop.

International

UK National Health Service (NHS)
The report advocates for an AI Navigation Assistant to streamline NHS triage, reduce inefficiencies, free up 29 million GP appointments annually, and improve patient outcomes by ensuring faster, more accurate care

Third Draft of the general-purpose AI Code of Practice
The third draft of the General-Purpose AI Code of Practice refines transparency, copyright, and risk mitigation commitments, with stakeholder feedback shaping its finalization as a compliance tool for the AI Act by May 2025.

Swedish gov proposes bill to allow police to use AI face-recognition
Sweden has proposed a law allowing police to use real-time AI-powered facial recognition to fight serious crime, aiming to curb gang violence while pledging to uphold personal privacy protections.

UK Tech Secretary to bang the drum for closer AI partnership with the US
UK Technology Secretary Peter Kyle is visiting the U.S. to promote AI investment and unveil Britain’s AI Growth Zones, positioning the UK as a global AI hub and strategic partner to the U.S.

AI is turbocharging organized crime, E.U. police agency warns
Europol warns AI is supercharging organized crime across Europe, enabling deepfakes, cyberattacks, and fraud that blur lines between criminal and state-sponsored threats.

Everything Else

The future of AI
The Future of AI: Perspectives for Startups report features insights from 23 AI leaders on key trends, innovations, and challenges for startups looking to scale in 2025.

AI-powered lawyering: AI reasoning models, RAG, and the future of legal practice
This study finds that AI tools, including RAG-powered applications and AI reasoning models, significantly enhance legal work quality and efficiency, suggesting a promising future for AI-assisted lawyering.

Why strategy beats speed in introducing AI for healthcare
Rushing AI adoption in healthcare risks failure and harm; a strategic, phased approach focused on readiness, alignment, and stakeholder engagement is essential to realizing AI’s full, equitable potential.

Perplexity AI in talks to double valuation to $18 billion
Perplexity AI is in talks to raise up to $1 billion at an $18 billion valuation, doubling its worth amid fierce AI search competition while facing controversy over content use.

Claude can now search the web
Anthropic’s Claude now delivers real-time, citation-backed answers by integrating web search, enhancing accuracy for tasks requiring up-to-date information.