2025 AI in Action Workshop Recap

Alexis Bonnell, former CIO of the Air Force Research Laboratory who now leads the U.S. National Labs Partnership at OpenAI, explores how human–machine teaming is driving breakthroughs in government innovation, science and national security.

Featured Speaker:

• Alexis Bonnell, Partnership Manager for the U.S. National Labs Partnership, OpenAI

As federal agencies expand their use of AI, questions about accountability, transparency, due process and legal oversight are becoming more urgent. This session explores how the courts are thinking about AI, such as how it could be used within judicial systems and how it might appear as part of legal challenges. Federal technologists would learn about the legal risks and ethical responsibilities shaping future case law and what that means for building AI tools.

Featured Speaker:

• Hon. Paul Grimm (ret.), Director,, Bolch Judicial Institute at Duke Law School

AI leaders are accelerating AI development and implementation to boost mission outcomes and ensure thoughtful innovation across agencies. Hear how experts are building on the foundation to transform federal operations, advance national priorities and stay ahead of emerging technological challenges.

Featured Speakers:

• Karen Howard, Executive Director, Office of Online Services, IRS
• Josh Phifer, Principal Solutions Architect, Elastic
• Steve Wallace, CTO and Director of Emerging Technology, DISA

As AI companies race to roll out chatbots across the federal government, agencies keep running into the same wall: “As an AI model, I cannot assist you with that task.” Glenn Parham, former generative‑AI lead at the Pentagon and founder of GovBench, breaks down the real blockers: policy gaps, benchmarking blind spots and the acute compute bottleneck. The session introduces GovBench’s AI + Government Alignment framework, mission‑specific benchmarks now in pilot and practical steps for labs, cloud providers, CIOs and CAIOs to coordinate so models can scale safely — and actually work — inside government.

Featured Speaker:

• Glenn Parham, CEO, GovBench

Joshua Powell, Chief AI Officer and chief innovation officer at Customs and Border Protection (CBP), outlined how CBP is responsible for over half of DHS’s AI inventory for border security, trade enforcement and traveler processing. He detailed how the agency operates computer vision that monitors thousands of remote cameras and large language models that cut administrative workflows in half. He also offered blunt advice to AI vendors: skip the generic slide decks and deliver relevant, tangible use cases that address real government needs.

Featured Speakers:

• Joshua Powell, CAIO and Chief Innovation Officer, CBP

Leveraging agentic AI solutions requires getting the basics right first. Julie Obenauer-Motley, senior responsible AI analyst at Johns Hopkins Applied Physics Laboratory, drew on her experience working within the defense sector to outline three lessons for success in AI: get the data right, ensure the platform works and involve users early. She shared how she has seen AI projects fail, like when data has not been digitized, systems without proper launchers and designs that fail in the real world.

Featured Speakers:

• Julie Obenauer-Motley Sr., RAI Analyst, Johns Hopkins University – Applied Physics Lab

Organizations today commonly approach automation in a linear view. This mindset can creep into how we build AI systems and can introduce siloed, high-maintenance and inefficient solutions. This session introduces the concept of multitask learning to enable AI models to perform multiple tasks simultaneously and help you move beyond narrow automation and toward systems that think in context.

Featured Speaker:

• Dr. Ali Naqvi, CEO, American Institute of Artificial Intelligence (AIAI)