AI Week in Review - 3/1/25

Public Sector

Author

Chris Kraft
March 01, 2025

Good morning, public sector AI enthusiasts. Welcome to the first edition of AI Week in Review, Public Sector, on our new platform. Thanks for making the move! During the month of March, I will keep both platforms updated. Starting in April, I will fully transition to this platform.

In this week’s edition:

  • Another week, another GenAI model

  • UC Berkeley releases GenAI Playbook

  • The latest public sector AI news

The Highlights

The Story: Anthropic just released Claude 3.7 Sonnet, their most intelligent model to date and the first hybrid reasoning model on the market.

More Details

  • Claude 3.7 Sonnet adds hybrid reasoning for faster, smarter responses.

  • Outperforms rivals in coding, excelling in SWE-bench and TAU-bench.

  • Available on all plans, API, Amazon Bedrock, and Google Cloud.

  • Pricing: $3M input / $15M output tokens.

  • Claude Code, a new AI coding tool, is in limited preview.

The Story: UC Brekeley just released a comprehensive AI Playbook. Created for product managers and organizational leaders, the Playbook outlines risks to be aware of when using GenAI plus 10 plays with accompanying resources and tools to use GenAI responsibly. 

More Details

  • GenAI is rapidly being adopted, offering productivity gains but requiring responsible use to unlock full benefits.

  • Key risks include data privacy, bias, transparency, inaccuracy, and security threats.

  • The playbook outlines 10 strategies, including governance frameworks, risk assessments, and leadership commitment.

  • Organizations that prioritize responsible AI gain a competitive edge, maintaining trust, compliance, and sustainable growth.

  • Developed with Google, the playbook provides real-world examples and tools to support responsible AI adoption.

The Details

Federal

Federal AI cuts risk U.S. innovation and global competitiveness
Trump’s NIST layoffs may weaken U.S. AI leadership, hinder CHIPS Act goals, and accelerate state-level AI regulations due to federal expertise gaps.

Chief AI officers must be preserved in the Trump administration
Chief AI officers (CAIOs) play a vital role in modernizing federal agencies, ensuring AI adoption aligns with efficiency, national security, and responsible governance—making them essential for the Trump administration’s AI strategy.

Federal budget cuts threaten to decimate America’s AI superiority
Proposed federal research cuts threaten America's AI dominance, undermining high-risk, high-reward innovations that drive economic growth, job creation, and technological leadership in a $15.7 trillion industry.

Senate passes AI deepfake bill, as Congress considers AI legislation
The Senate unanimously passed the TAKE IT DOWN Act criminalizing non-consensual AI deepfakes, while Congress considers 13 AI-related bills addressing security, trade, healthcare, and ethical concerns.

Anthropic partners with U.S. National Labs for first 1,000 Scientist AI Jam
Anthropic joins the Department of Energy’s first-ever AI Jam, where 1,000 scientists will evaluate AI’s ability to accelerate research, solve complex problems, and enhance national security applications.

Space Force turns to AI
As space grows more congested, the Space Force is exploring AI to enhance space domain awareness, automate monitoring, and reduce reliance on human operators for critical decision-making.

Energy Secretary compares current AI development to the Manhattan Project
U.S. Energy Secretary Chris Wright likened AI development to the Manhattan Project, emphasizing its impact on national security, energy demand, and technological competition while advocating for nuclear energy expansion.

The real threat of Chinese AI
China's DeepSeek AI models are disrupting the global AI landscape with powerful, low-cost, open-source alternatives, challenging U.S. dominance and raising concerns over Chinese influence in AI development and infrastructure.

Senate bill targets DeepSeek ban from government devices
A bipartisan Senate bill seeks to ban China's AI platform DeepSeek from all federal government devices, citing security concerns and its ties to the Chinese government.

State / Local

Arizona’s release AI election security report
Arizona’s AI and Election Security Advisory Committee urges the creation of an AI learning lab, enhanced cyber defenses, and collaboration with tech firms to protect elections from AI-driven threats.

MTA and Google public sector tackle preventive track maintenance
The MTA's TrackInspect pilot uses AI, cloud technology, and real-time sensor data from subway trains to detect track defects early, improving safety, reducing delays, and optimizing maintenance.

New York expands ‘AI for good’ research consortium
Gov. Kathy Hochul announced expanded funding and participants for Empire AI, a state-backed consortium using AI to tackle public challenges like healthcare, climate change, and cybersecurity.

Kansas Supreme Court forms AI committee
The Kansas Supreme Court established a 21-member committee to study AI’s role in the legal system, ensuring responsible use while improving court efficiency and accessibility.

Indianapolis teaches ethical AI to government workers
After a survey revealed low AI policy awareness, Indianapolis and Marion County partnered with InnovateUS to provide AI ethics training and data security education for public employees.

Delaware State University adds AI gun detection
Delaware State University has implemented ZeroEyes’ AI-powered gun detection system, which integrates with existing cameras and provides rapid alerts to law enforcement, enhancing campus security.

Georgia General Assembly examines AI
The Georgia General Assembly is reviewing multiple AI-related bills, with a focus on healthcare advancements, ethical concerns, and statewide innovation while balancing safeguards and regulation.

AI to help Catoosa County track tax exemptions
Catoosa County is adopting an AI-powered system to identify ineligible homestead exemptions and help eligible residents claim tax benefits, aiming to ensure fairness and accuracy in property taxation.

K-State launches AI chatbot that checks in on students
Kansas State University has launched "Willie," an AI-powered chatbot that texts students throughout the semester to check on their well-being and connect them with resources.

Md. lawmakers propose AI cameras to combat distracted driving, raising privacy concerns
Maryland lawmakers are considering AI-powered cameras to detect distracted driving, raising privacy concerns while aiming to reduce crashes and fatalities with a proposed $40 civil fine.

Nevada judiciary launches enhanced self-help legal website with AI support
Nevada’s judicial branch has expanded its legal self-help website with AI-powered multilingual support and step-by-step guides to assist self-represented litigants in navigating the legal system.

Apple to build AI server factory in Texas
Apple is investing $500 billion in U.S. expansion, including a new AI server facility in Houston, creating 20,000 jobs and strengthening its commitment to American manufacturing.

Expelled student sues U. Minnesota after claims of AI use
A former Ph.D. student is suing the University of Minnesota for over $1 million, alleging wrongful expulsion after professors accused him of using AI to write essays.

International

Overview of AI policy in 15 jurisdictions
AI policies across 15 jurisdictions reflect a global shift toward balancing innovation with ethical and societal considerations, promoting international cooperation for responsible AI development.

AI safety must reflect Caribbean voices, culture, and sovereignty
Global AI safety frameworks often ignore the Caribbean’s cultural and linguistic diversity, risking misrepresentation—inclusive policies must ensure these regions have a voice in AI governance.

India shaping AI governance policies
India is advancing AI governance and innovation, with 67 proposals for indigenous AI models and a mission to build ethical, inclusive, and trustworthy AI.

Taiwan uses AI to combat online fraud and deep fakes
Taiwan’s MODA leverages AI and machine learning to combat cyber fraud, enhance digital governance, and drive nationwide digital transformation through smart government initiatives and data-driven innovation.

UK delays plans to regulate AI as ministers seek to align with Trump administration
The UK postpones AI regulation, prioritizing investment appeal over immediate oversight, as it seeks closer alignment with the Trump administration’s approach.

Ireland charts a responsible AI future with strategic policy recommendations
Ireland’s AI Advisory Council outlines strategic recommendations to drive AI adoption, strengthen competitiveness, and ensure ethical, sustainable AI development across key sectors.

Hong Kong to cut thousands of civil service jobs and invest in AI to tackle a rising deficit
Facing an $11.2 billion deficit, Hong Kong will reduce government spending by 7%, cut 10,000 civil service jobs, freeze salaries, and boost AI investment to restore fiscal balance.

Can AI help modernize Ireland's healthcare system?
Irish hospitals are adopting AI for faster diagnostics and patient care, but outdated IT systems and regulatory challenges could slow the technology’s full potential in healthcare modernization.

Canada launches probe into X’s AI data collection
Canada’s privacy commissioner is probing X’s handling of user data for AI training, following a complaint about potential violations of federal privacy laws.

Stuck in eternal drought, UAE turns to AI to make it rain
The UAE is leveraging AI to enhance cloud seeding and improve rainfall predictions, but experts caution that data limitations and human oversight remain crucial for success.

Japan Launches National AI Librarian Project
Japan's National AI Librarian Project aims to boost AI literacy among students and professionals through practical training, helping bridge skill gaps and enhance national competitiveness in AI technology.

Defense Ministry to test out GenAI in South Korea-US military drill
South Korea’s Defense Ministry will test its new generative AI system, GeDAI, during next month’s joint military drill with the U.S., marking the first use of generative AI in allied exercises.

How Russia is losing the AI race
Russia lags behind in generative AI due to hardware shortages, brain drain, and strict censorship, preventing it from competing with leading models like ChatGPT and DeepSeek.

Indonesia advocates for ethical AI
Indonesia's Deputy Minister of Communication and Digital Affairs stresses the importance of ethical AI development, emphasizing transparency, accountability, inclusivity, and data protection.

Africa navigates challenges to bridge the AI divide
Sub-Saharan Africa faces major barriers to AI adoption, including poor digital infrastructure, high illiteracy rates, and brain drain, but partnerships and investment could help close the AI divide.

Everything Else

Autonomous AI: Powerful, promising, and potentially perilous
AI agents promise efficiency but pose risks like cascading errors, bias, and security threats, requiring careful oversight.

How disruptive is DeepSeek? Stanford faculty weigh in
China’s DeepSeek AI model challenges U.S. dominance, sparking debates on open-source innovation, geopolitical risks, and AI’s impact on research, regulation, and global markets.

UC Berkeley guide helps business leaders navigate AI ethics amid rapid adoption
This GenAI playbook outlines key risks and strategies for responsible generative AI use, emphasizing transparency, governance, and ethical implementation for long-term trust and sustainability.

Anthropic releases Claude 3.7
Claude 3.7 Sonnet, a hybrid AI model combining real-time responses with deep reasoning, aiming to simplify AI interactions and gain an edge over rivals.

Anthropic finalizing $3.5 billion funding round
Investor confidence in AI remains strong as Anthropic raises $3.5 billion at a $61.5 billion valuation, positioning itself to compete with OpenAI despite rising competition from China’s DeepSeek.

How Ikea approaches AI governance
Ikea integrates AI governance into business priorities, ensuring responsible use, compliance with the EU AI Act, and a proactive approach to ethical AI development.

American workers are skeptical AI will help them on the job
A Pew survey finds most American workers rarely use AI on the job and are more worried than excited about its future impact, fearing job displacement over productivity gains.

Can AI predict which cancer patients should be treated with immunotherapy?
Researchers have developed SCORPIO, an AI-based tool that uses routine blood tests and medical records to better predict patient response to immune checkpoint inhibitors, surpassing FDA-approved tests in accuracy.

Meta plans standalone AI app to compete with ChatGPT
Meta plans to release a standalone AI assistant app in Q2, expanding Meta AI beyond its existing platforms, with a potential paid subscription model to compete with OpenAI and Google.

OpenAI unveils GPT-4.5, its largest AI model yet
OpenAI launches GPT-4.5, a powerful but costly AI model with improved accuracy and emotional intelligence, available to ChatGPT Pro users and API developers at a premium price.

AI job market surges 68% as demand for AI skills reshapes the workforce
AI-related job postings have skyrocketed since ChatGPT’s launch, replacing traditional IT roles, while companies invest heavily in AI-driven efficiency and automation, fueling concerns over job displacement.

IT leaders prioritize AI for productivity gains over innovation
A new survey reveals that IT executives are focusing AI investments on efficiency and workforce reductions rather than innovation, while debates continue over AI pricing and long-term value.

Google faces lawsuit over AI search summaries as businesses report lost traffic
Chegg is suing Google, claiming AI-generated search summaries divert traffic and harm revenue, raising broader concerns about AI’s impact on SEO and content creators.

Raiders GM John Spytek drops huge take on AI scouting
New Raiders GM John Spytek is exploring AI to enhance scouting and decision-making, signaling a tech-driven approach to rebuilding the team's roster.