The Office of Management and Budget (OMB) recently released a memo that outlines how federal agencies should approach purchasing artificial intelligence (AI) technologies. This memo doesn't introduce groundbreaking ideas or repeat past guidance. Instead, it acknowledges the ever-evolving nature of AI acquisition and usage within government agencies.

Emphasizing Oversight in Contracts

Kraig Conrad, who leads the National Contract Management Association, highlighted the memo's focus on maintaining oversight after contracts are awarded. He pointed out that post-award phases are often where financial inefficiencies occur. Conrad emphasized the importance of regular check-ins and reporting requirements to ensure projects stay on track and deliver value. This approach aims to prevent the loss of contract value and ensure that someone is actively monitoring project progress.

Deadlines and Responsibilities

The memo sets out four key deadlines. The General Services Administration (GSA) has two: within 100 days, it must create a public guide for agencies on purchasing AI tools, and within 200 days, it should establish an online portal for federal agencies to share knowledge and resources related to AI procurement.

Agencies also face two deadlines. In 270 days, they need to update acquisition procedures to align with the memo, including reviewing AI-related acquisitions and forming cross-functional teams to aid decision-making. In 200 days, agencies must establish processes and contract terms to manage government data use and clarify intellectual property rights.

Addressing Implementation Challenges

While the memo received positive feedback, experts like Conrad note that many questions about implementation remain. One industry insider mentioned that some deadlines are set for fiscal 2026, leaving current AI purchases unaddressed. The memo builds on previous guidance from the Biden administration, emphasizing avoiding vendor lock-in and maintaining data portability and ownership.

John Ferry, president of Trenchant Analytics, praised the memo's focus on preventing vendor lock-in, a common issue when government data is modified and becomes proprietary to companies. However, he acknowledged that agencies often struggle with data management, a crucial step before tackling intellectual property concerns.

Shifting Contracting Approaches

A notable difference in the memo is its recommendation for agencies to use specific contract types when acquiring AI tools. It suggests focusing on statements of objectives and performance-based contracts with incentives, rather than fixed-price contracts. This approach encourages agility and learning, allowing agencies to adapt to new features and technologies.

Ferry and other experts agree that this shift may be challenging for some agencies. The complexity of these contracts and the need for frequent updates could pose difficulties, especially as experienced contracting officers retire.

AI as a Solution for Contracting Burdens

The memo also encourages using AI tools to streamline acquisition processes. For instance, the Department of Homeland Security has piloted AI for market research and contract closeouts. The GSA is testing generative AI tools to assist contracting officers in navigating acquisition policies.

Ferry sees AI as a perfect fit for acquisition processes due to their standardized and linear nature. With pressures mounting on the acquisition workforce, AI offers a promising solution to fill gaps and enhance efficiency. The workforce is eager to adopt AI, and the ethical concerns are minimal compared to other government applications.

In summary, the OMB's memo provides a framework for federal agencies to navigate the complexities of AI procurement. While challenges remain, the emphasis on oversight, flexible contracting, and AI integration offers a path forward for effective and responsible AI adoption in government.