AI in the Public Sector

The Numbers Are Clear

Here’s the state of AI in government right now: federal agencies reported over 1,100 active AI use cases in 2024 — a ninefold increase in just one year. Over 90% of states have adopted responsible AI policies. In New Jersey’s AI sandbox, 20% of state workers actively use AI weekly, saving hours on routine tasks.

This isn’t future technology. It’s current practice. And the question isn’t whether your agency will adopt AI, but whether you’ll be ready when it does.

What You’ll Learn

Objectives for this course:

Apply AI to draft reports, memos, and public communications
Use AI for data analysis and policy research
Evaluate AI outputs for bias, accuracy, and ethics compliance
Implement AI workflows for constituent services and case management
Design responsible AI governance aligned with NIST frameworks
Create emergency communications and crisis documentation with AI

How This Course Works

Each lesson tackles one aspect of government AI use:

Lesson	Focus	What You’ll Do
1	Introduction	Assess your AI readiness and data landscape
2	Writing	Draft reports, memos, and public communications
3	Data & Policy	Analyze data and research policy impacts
4	Constituents	Handle inquiries, case management, FOIA
5	Ethics	Prevent bias and ensure responsible use
6	Emergencies	Draft crisis communications and response plans
7	Procurement	Write RFPs, justify budgets, ensure compliance
8	Full Toolkit	Build your complete AI workflow

What to expect: Each lesson takes 12-18 minutes. Exercises use real government scenarios. Every AI output in this course goes through human review — because that’s how government AI should work.

Where AI Helps Most in Government

Based on the Federal AI Use Case Inventory and research from Harvard’s Kennedy School and Ash Center, here are the highest-impact areas:

Task	Time Without AI	Time With AI	What AI Does
Draft a 5-page report	6-8 hours	2-3 hours	First draft, structure, research summary
Process FOIA request	Days to weeks	Hours to days	Document search, relevance filtering, redaction flagging
Prepare meeting minutes	2-3 hours	30-45 minutes	Transcription, action item extraction, formatting
Analyze policy impact	1-2 weeks	2-3 days	Data synthesis, comparison, stakeholder analysis
Respond to constituent inquiry	30-60 minutes	10-15 minutes	Draft response, FAQ matching, routing

✅ Quick Check: Why must AI outputs in government always go through human review?

Because government decisions carry legal weight and affect real people’s lives. An AI-drafted denial of benefits, if sent without review, could violate due process. An AI-summarized policy document, if it misses a critical exception, could lead to incorrect enforcement. Government accountability requires a human who understands the context, can verify accuracy, and can be held responsible for the final decision. AI assists; humans decide.

Your First Step: AI Readiness Assessment

Help me assess my agency's AI readiness:

My role: [your title and department]
Agency level: [federal / state / local]
Current AI policy: [we have one / don't know / none exists]
Biggest time sink in my work: [what tasks eat your hours?]
Data I work with: [public records, constituent data, financial data, personnel, etc.]
Current tools: [what software/systems do you use daily?]

Assess:
1. Which of my daily tasks could AI assist with right now?
2. What data classification issues should I check before using AI?
3. Does my agency likely have an AI use policy? Where would I find it?
4. What's the lowest-risk, highest-impact AI task I could start with?
5. What training resources exist for government employees at my level?

Be conservative — flag any tasks where AI use would need IT security approval first.

Data Classification: Know Before You Go

Before using any AI tool, classify the data you’d be inputting:

Classification	Can Use With AI?	Examples
Public	Yes (most tools)	Published reports, press releases, meeting agendas
Sensitive but unclassified	Agency-approved tools only	Constituent names, internal memos, draft policies
Personally identifiable (PII)	Approved tools with data agreements	Social security numbers, addresses, case files
Classified	Never with commercial AI	National security, law enforcement, intelligence

Rule of thumb: If you wouldn’t email it to a stranger, don’t put it into a consumer AI tool. Use your agency’s approved platform instead.

Key Takeaways

Government AI use has exploded: 1,100+ federal use cases, 90%+ states with AI policies, ninefold GenAI increase in one year
AI saves government workers hours per week on drafting, summarizing, and data processing
Every AI output in government must go through human review — accountability requires it
Check data classification before using any AI tool — public data is generally safe; PII and sensitive data require approved platforms
Start with low-risk, high-impact tasks: meeting preparation, report drafting, research summarization

Up Next: In the next lesson, you’ll learn to write government reports, memos, and public communications with AI — including how to meet plain language requirements and maintain your agency’s voice.