Responsible AI for Public Organizations
Designing AI Projects in Government - Resources
Find all the resources mentioned in the online course by selecting a module from the dropdown menu below. Have questions? Contact us
Ready for a deeper dive? Take a course to
Module 2: Overview of AI approaches
This module provides a comprehensive introduction to artificial intelligence (AI), machine learning, and generative AI. It explores the three main approaches to machine learning (supervised learning, unsupervised learning, and reinforcement learning) and the crucial decision of choosing the right algorithm for an AI project. The module also discusses the differences between generative AI and traditional machine learning, highlighting the accessibility and challenges of generative AI.
Learn more
-
Prof Narayan’s (author of AI Snakeoil) guide to thinking through what tasks more complicated neural networks are good for: How to Recognize AI Snake Oil Professor Arvind Narayanan (video, slides) [Archive]
-
PAIR Guide Ch. 4 Explainability & Trust from Google - while written w/ private sector companies in mind, provide good overview of thinking through explainability and trust in AI systems
-
Different Approaches - Generative AI
-
An accessible animated explainer of how generative works under the hood from
-
A more technical interactive explainer of how generative AI predicts the next word from PAIR (People, AI, Research) group at Google.
-
IBM Blog Post + Video - What is Generative AI
Citations
-
InnovateUS Course: Responsible AI for Public Professionals: Using GenAI at Work - https://bit.ly/genai-at-work
-
Washington Post - Is artificial intelligence about to transform the mammogram? - https://wapo.st/3UZF5hV
-
MIT News - Using AI to predict breast cancer and personalize care - https://bit.ly/mit-ai-cancer
- Public Safety Assessment (PSA) tool - https://bit.ly/psa-tool
Module 3: Surfacing and selecting AI projects for scale
This module focuses on identifying and prioritizing AI projects that deliver meaningful results for the organization and the public. It presents three methods for recognizing business challenges that could benefit from AI solutions. The module also explores the criteria for evaluating and selecting AI projects and the pivotal role of pilot programs in validating AI projects before wider-scale deployments.
Learn more
-
Problem Definition
-
InnovateUS Course: Innovation Skills Accelerator - Module 2 Introduction to Problem Definition
-
InnovateUS Course: Human Centered Design - Module 3 Define
-
Surfacing and selecting projects - Common problem type or ‘typology’
-
Blog series from DataSF on how they run their DataScienceSF program (Part 1, Part 2, Part 3, Part 4) [Archive]
-
Blog by Nolaytics on their problem typology (blog) [Archive]
- InnovateUS Workshop: Delivering Data Scient Projects in Government by Joy Bonaguro (video link)
-
Surfacing and selecting projects - Bottom up experimentation and crowd sourcing
-
InnovateUS Course: Innovation Skills Accelerator - Module 6 Collective Intelligence: co-creation, collaboration, and crowdsourcing
-
Surfacing and selecting projects - Fast field scanning
-
InnovateUS Course: What Works: Fast Field Scanning
-
Lists of government use cases
-
HHS AI use cases [Archive]
-
DOL AI use cases [Archive]
-
Commerce AI use cases [Archive]
-
Pilots
-
InnovateUS Workshop: Effective Measurement and Pilots by Carin Clary & Kate Lawyer
Citations
-
InnovateUS Course: Innovation Skills Accelerator https://bit.ly/skills-accel-course
-
InnovateUS Course: Human Centered Design https://bit.ly/hcd-course
-
Nolaytics - Types of analytics projects https://bit.ly/nolaytics
-
DataSF - Data Science Typology https://bit.ly/datasciencesf
-
DataSF - https://bit.ly/datasf
-
Nolaytic - https://bit.ly/nolaytics
-
InnovateUS Course: Responsible AI for Public Professionals: GenAI at Work https://bit.ly/genai-at-work
-
InnovateUS Course: What Works: Fast Field Scanning https://bit.ly/ffs-course
-
AlertCalifornia - https://alertcalifornia.org