Artificial Intelligence Training Series for Federal Employees
By the Artificial Intelligence Community of Practice
The emergence of artificial intelligence (AI) is being heralded as among the most consequential technological advances of the 21st century. AI and related technologies are already transforming economies, challenging legal and political norms, and reconfiguring society. Government entities attempting to navigate this new set of technologies must adapt to better serve the American people
The Stanford University Institute for Human-Centered AI (HAI) has tailored a government education program specifically for the U.S. General Services Administration’s AI Community of Practice (AICoP), in partnership the Office of Management and Budget’s Office of the Federal Chief Information Officer (OMB OFCIO). This program will explore the latest in AI developments, equipping participants with knowledge needed to think critically about implementing and governing this emerging technology.
About This Program
Understanding the impact of AI on society is a multi-faceted enterprise that requires expertise from computer science, economics, law, political science, psychology, and a host of other disciplines. This program draws upon the knowledge of multidisciplinary AI experts in academia, as well as leaders and case studies from civil society and industry.
This program is offered by Stanford HAI, in partnership with OMB-OFCIO and the AI CoP. These sessions are intended to introduce concepts and theory around artificial intelligence to empower federal employees with awareness of the technical concepts underlying AI, including various algorithm structures in AI and the technical underpinnings of foundation models, as well as examples of risks that have occurred through the use of AI and techniques that AI developers can use to mitigate risk. Per the requirements as laid out by the AI Training Act, attendees should expect the session content to include technical aspects of AI systems and theory. However, the views presented in this session are not representative of any formal government policy, nor should they be interpreted as official guidance to the use of AI by federal agencies. Please consult OMB’s Information and Guidance page on WH.gov to reference existing policy.
These trainings are open to all federal employees, though there is specific emphasis for the acquisition workforce, program managers, budget professionals, and staff involved in creating federal systems. Registration is required for each session via Zoom with a .GOV or .MIL email address.
Program Schedule
Navigating the AI Landscape: Fundamentals of the Science and Technology
Monday, September 18 | 12:00pm-1:00pm ET | Peter Norvig
This introductory session will provide a comprehensive overview of AI, including the definition, theories of AI and machine learning, neural networks, narrow vs. general AI, gradient descent, use cases, and more. It will provide a bird’s-eye view of the AI landscape, covering different AI techniques such as deep learning, computer vision, natural language processing, and supervised and unsupervised learning.
Participants will gain a foundation on the central concepts, techniques, and applications that comprise the exciting and rapidly evolving field of AI.
Understanding Foundation Models: Opportunities and Challenges
Thursday, September 21 | 12:00pm-1:00pm ET | Rishi Bommasani
Recently, a new paradigm for building AI systems that powers many generative AI systems has emerged: foundation models that were trained on huge amounts of data and can be adapted to numerous applications. Foundation models are at the center of the global discourse on AI: the emerging technological paradigm has a concrete and growing impact on the economy, policy, and society.
Session two explores foundation models — how they are created and deployed, the requirements to build one, the expected and unexpected consequences of these models, and the evolving landscape of regulations, guidelines, and best practices with insights into key considerations for effective governance, including transparency, accountability, and inclusivity.
Mitigating Risk: Implementing Safe & Robust AI
Friday, September 22 | 12:00pm-1:00pm ET | Anthony Corso
The consequences of deploying robust AI and decision-making technologies in safety-critical systems such as driverless vehicles and autonomous aircraft are enormous. Challenges for AI developers range from biased inputs, to constantly evolving conditions, and explainability issues, among others.
Session three will discuss the obstacles developers face as well as the difficult—and often politically fraught—decisions they make around operational efficiency and how they define acceptable risk parameters.
Modernizing a Mammoth: Use-Cases of Public Sector AI
Tuesday, September 26 | 12:00pm-1:00pm ET | Daniel E. Ho
From streamlining administrative processes to providing personalized services to constituents, there is ample opportunity for AI to help government agencies achieve their missions. However, integrating AI into the government is not as easy as obtaining and deploying the technology. Talent, infrastructure, public trust, and morale play equally important roles in ensuring the successful modernization in government organizations.
Session five will dive into current use cases of AI in government, the challenges and successes of these cases, and how to improve the integration of new technologies that will help the government serve its citizens.
Addressing Bias and Data Privacy Concerns in the Age of Algorithms
Wednesday, September 27 | 12:00pm-1:00pm ET | Jen King
AI designers face enormous challenges in acquiring data, training AI technologies, and addressing the biases in data (and their non-obvious proxies) within their AI systems. Moreover, different social values around privacy, data ownership, and data creation will impact what AI technologies are possible today and what the future paths of innovation in AI will look like.
Session four will unpack key ethical questions on bias and discrimination in AI systems and different social values around privacy, data ownership, and data creation impacting AI.
The Future of AI Innovation: A North Star for Robotic Learning and Embodied AI
Thursday, September 28 | 12:00pm-1:00pm ET| Dr. Fei Fei Li
AI technologies are reshaping the traditional notions of human-computer interaction. How can we build AI to augment human capabilities, not replace them?
Session six will explore the future of AI innovation, particularly the transformative power of robotic learning and embodied AI. Drawing inspiration from advancements in computer vision and natural language processing, participants will explore AI’s potential in understanding and engaging with diverse environments.
About the AI CoP
Communities of practice are government-wide interest groups that provide opportunities for members to support, challenge and inspire one another. Cross-disciplinary and cross-agency by design, the goal of the communities is to expand your knowledge and practice in a given area of interest. The Artificial Intelligence Community of Practice (AI CoP) unites federal employees who are active or interested in AI policy, technology, standards, and programs to accelerate the thoughtful adoption of AI across the federal government. Learn more and join the community here.
*This training is only available to government employees. This registration will grant access to all sessions. Participation in all sessions is encouraged, but not required.
Return to Updates