AI Week in Review - 4/5/25

Hello, public sector AI enthusiasts. Lots of activity last week across federal and state/local agencies.

In this week’s edition:

• White House Releases AI Guidance

• General Public versus AI Experts

• Agent Companions

• Latest public (and private) sector news

The Highlights

White House Releases AI Guidance

The Story
The White House released updated versions of OMB M-24-10 (Advancing Governance, Innovation, and Risk Management for Agency Use of Artificial Intelligence) and M-24-18 (Advancing the Responsible Acquisition of Artificial Intelligence in Government).

More Details
OMB M-25-21 (Accelerating Federal Use of AI through Innovation, Governance, and Public Trust) rescinds and replaces M-24-10. This memo directs agencies to provide improved services to the public, while maintaining strong safeguards for civil rights, civil liberties, and privacy.

• Driving AI Innovation: Creating an AI Strategy, sharing data/AI, buying American, AI procurement (pointing to M-25-22), and preparing an AI-ready workforce.

• Improving AI Governance: CAIO, AI Governance Boards, agency governance responsibilities, and federal governance responsibilities.

• Fostering Public Trust in Federal AI: Identify high-impact AI and ensure minimum practices are met for high-impact.

OMB M-25-22 (Driving Efficient Acquisition of Artificial Intelligence in Government) rescinds and replaces M-24-18. This memo provides guidance to federal agencies for them to improve their ability to acquire AI responsibly.

The three themes of M-25-22 are:

• Ensuring the Government and the Public Benefit from a Competitive American Al Marketplace

• Safeguarding Taxpayer Dollars by Tracking AI Performance and Managing Risk

• Promoting Effective AI Acquisition with Cross-Functional Engagement

General Public versus AI Experts

The Story
The Pew Research Center looked at how the general public and AI experts have different enthusiasm and predictions for AI. One thing they found was that AI experts have a more positive outlook on AI versus the general public.

More Details

Here are a few of the observations:

• Experts are far more positive and enthusiastic about AI than the public

• Larger shares of experts than of U.S. adults see AI as personally beneficial.

• Public optimism is low regarding AI’s impact on work.

Google Goes Agentic

The Story
Google released an Agent Companion whitepaper that covers agentic AI. This is a “102” level guide that’s designed for developers. The original 2024 Google Agent paper was released earlier this year.

More Details
Agents have been a hot topic for some time and these resources help orient users on how to move forward with agentic workflows.

The Details

Federal

Trump White House releases guidance for AI use, acquisition in government
The Trump administration released new AI guidance for federal agencies that replaces Biden-era directives but retains key structures—emphasizing innovation, risk management, and a push for U.S.-made AI technologies.

Senators open investigation into DOGE’s Dept of Ed Chatbot
Senators Warren, Schumer, and Hirono launched an investigation into DOGE’s plan to replace student aid call centers with AI, citing risks of misinformation, privacy breaches, and conflicts of interest.

House Republicans forecast ‘light touch’ regulatory approach to AI under Trump
House Republicans pledge minimal AI regulation under Trump, while Democrats warn of risks tied to DOGE’s access to federal data and Elon Musk’s growing influence over government AI systems.

Cyber Command touts AI-driven gains in cybersecurity, network monitoring
Cyber Command reports major cybersecurity gains using generative AI, cutting threat analysis time from days to minutes and enhancing real-time network monitoring across the Department of Defense's global infrastructure.

Energy invites AI data center development at Los Alamos and other federal lands
The U.S. Energy Department is inviting tech firms to build AI data centers on 16 federal sites, highlighting nuclear-powered infrastructure to fast-track development amid surging demand for AI computing capacity.

Will Congress’s new AI plan survive Trump?
The bipartisan ASAP initiative aims to supercharge U.S. science with AI, but its future hangs in the balance as Trump-era budget cuts clash with industry-backed ambitions for tech-driven research acceleration.

Career experts urge federal workers to learn AI skills as Trump cuts workforce
As federal layoffs surge under Trump, career experts urge government workers to reskill in AI and data management to remain competitive and transition successfully into the evolving private sector.

Could a FedRAMP rehaul usher in the US government’s AI age?
As FedRAMP eyes a major overhaul, AI companies await streamlined authorization processes that could unlock broader access to the U.S. government—potentially accelerating public-sector AI adoption at scale.

AI screening for opioid use disorder associated with fewer hospital readmissions
An NIH-funded clinical trial found AI screening for opioid use disorder matched provider-led referrals in effectiveness and significantly reduced hospital readmissions, highlighting AI’s real-world potential to improve addiction care and cut costs.

AI for military decision-making
Military leaders are urged to balance AI's potential for faster, smarter decision-making with clear scopes, quality data, and human oversight to mitigate risks like bias and misuse.

The macroeconomic effects of AI
Congressional Research Service outlines how AI may impact the U.S. economy through productivity, labor markets, and inequality, though widespread economic transformation will likely take years.

State / Local

New Jersey governor signs bill criminalizing deepfakes
New Jersey has enacted a landmark law criminalizing malicious deepfakes, aiming to protect citizens from AI-driven deception in harassment and elections, joining a growing wave of state-level regulation.

New York needs better AI governance
A state audit warns New York lacks sufficient AI oversight, urging stronger policies, training, and inventories to prevent bias and misuse—just as the state’s first chief AI officer begins work.

Denver to create prequalified AI vendor list
Denver issued an RFP to create a prequalified pool of AI vendors, aiming to fast-track secure, innovative solutions that improve city operations and resident services while aligning with strategic goals.

State efficiency bill would look to leverage AI
North Carolina’s proposed DAVE Act would use AI to assess state agency performance and spending, raising bipartisan concerns about workforce impacts, transparency, and the growing role of algorithms in governance.

Northeastern Univ launches Claude for Education
Northeastern University has launched "Claude for Education," an AI tool by Anthropic designed to foster student critical thinking and support faculty, marking a major step in AI integration in higher ed.

Culver City uses AI to enhance street safety
Culver City is deploying AI-powered cameras to deter double parking in bus lanes, aiming to boost transit efficiency and safety while ensuring human oversight of ticketing decisions.

New traffic study suggests AI tech could improve traffic
A Florida study found outdated traffic signals cause longer wait times; researchers recommend AI-powered lights to optimize flow, reduce delays, and enhance safety—offering a cost-effective alternative to road expansion.

Law enforcement uses AI to stop flow of fireworks
Hawaii is using AI to identify patterns in illegal fireworks imports, aiming to curb smuggling at ports and compensate for staffing shortages following a deadly explosion tied to illicit fireworks.

Illinois exploring speed cameras, AI to reduce speeding
Illinois lawmakers are considering speed and AI camera technology to curb crashes on DuSable Lake Shore Drive, commissioning a study to evaluate its effectiveness in reducing accidents and improving driver behavior.

AI data center boom comes to the heartland
Meta’s $10B AI data center in rural Louisiana highlights a growing trend of tech expansion into small towns—offering economic transformation but raising concerns about long-term energy costs for residents.

AI cameras catch tens of thousands ahead of Distracted Driving Awareness Month
AI cameras in Minnesota detected over 10,000 distracted driving incidents in one month, prompting expanded enforcement and new tech deployments during Distracted Driving Awareness Month to curb rising crash fatalities.

Pennsylvania just demolished its largest coal plant to make way for a $10 billion AI data center
A former coal plant in Homer City, Pennsylvania, is being redeveloped into a $10B+ natural gas-powered AI data center, reflecting a growing trend of repurposing legacy energy sites for tech infrastructure.

Schools selected for new AI pilot program
Two Lowcountry middle schools were selected for South Carolina’s Palmetto AI Pathways pilot, giving students early exposure to AI and robotics through hands-on STEM learning and industry partnerships.

AI used to ID mosquitos, control diseases
Utah is developing AI tools to identify mosquito species faster and more accurately, helping monitor disease vectors like West Nile Virus and insecticide resistance across hybrid populations.

Virginial law puts guardrails on AI in customer support
Virginia’s new law regulates “high-risk” AI in customer service, requiring transparency, bias safeguards, and customer notifications—backed by fines up to $10,000 per violation starting July 2026.

International

Police partner with AI in arms race against criminals
Police across Canada are increasingly using AI to combat crimes like child exploitation, but experts warn the rapid adoption risks privacy violations, bias, and a lack of legal oversight.

Singapore releases agentic AI primer
Singapore’s Government Technology Agency releases a primer unpacking Agentic AI, offering public service stakeholders a clear guide to its capabilities, use cases, and implementation strategies beyond traditional LLMs.

India rides the agentic AI wave with Over 80% exploring potential
A Deloitte report finds over 80% of Indian businesses are embracing Agentic AI, with growing investment in GenAI automation—though scaling challenges, tool adaptability, and ROI concerns remain significant hurdles.

Rebooting copyright: How the UK can be a global leader in the arts and AI
The Tony Blair Institute urges the UK to lead in AI and the arts through balanced copyright reform, proposing a text and data mining exception with opt-out to unlock creative and economic value.

Governing in the age of AI: Unlocking a transformation in Africa
Africa must embrace AI as a strategic tool for transformative governance, using it to overcome systemic challenges, accelerate development, and leapfrog traditional models despite fiscal constraints.

Could Canada lead the world in AI regulation?
As the U.S. retreats from AI oversight, experts argue Canada has a unique opportunity to lead globally by enacting comprehensive AI regulation focused on privacy, innovation, and cybersecurity.

Everything Else

Everyone’s talking about AI agents. Barely anyone knows what they are
Despite the hype, tech leaders admit there’s no clear consensus on what qualifies as an AI agent, leading to confusion and mislabeling across the enterprise sector.

Google releases agentic AI companion
This paper dives into the architecture, orchestration, and real-world deployment of generative AI agents, spotlighting Google's tools and a multi-agent automotive case study to explore production-grade agent design.

How the U.S. public and AI experts view AI
A new Pew study reveals a stark divide between AI experts and the U.S. public on AI optimism, but both groups align in desiring more personal control and stronger government oversight.

Papa Johns wants AI to transform pizza ordering
Papa John’s is partnering with Google Cloud to personalize pizza ordering through AI-powered chatbots, virtual assistants, and targeted marketing, aiming to reclaim its tech leadership and boost customer engagement.

Being human in the age of AI
A global expert survey predicts AI will profoundly reshape human cognition, emotion, identity, and relationships by 2035—raising concerns over mental well-being, agency, and trust in cultural norms.

Sovereign remedies: Between AI autonomy and control
As global AI development accelerates, countries are increasingly adopting “sovereign AI” strategies—balancing legality, economic value, security, and cultural alignment to retain control over AI systems and their societal impact.

Cyberattacks by AI agents are coming
As AI agents grow more capable, experts warn of their potential to automate cyberattacks at scale—prompting urgent efforts to detect, benchmark, and defend against this emerging security threat.

MBA students at Harvard Business School must take AI course
Harvard Business School now requires all MBA students to take an AI-focused course, positioning them for a competitive edge as AI becomes central to modern business leadership and decision-making.

AI could affect 40% of jobs and widen inequality gaps
A UN report warns AI could disrupt 40% of jobs and deepen global inequality, urging inclusive governance, open-source tools, and international cooperation to ensure shared benefits from AI-driven growth.

OpenAI gets fresh funding — and momentum
OpenAI has secured $40 billion in fresh funding from SoftBank, boosting its valuation to $300 billion as it expands image tools, teases hardware, and pushes forward despite structural and legal challenges.