The White House is pushing back on a Government Accountability Office recommendation regarding how the executive branch should communicate with federal agencies about artificial intelligence governance, as revealed in a recent report from the congressional watchdog.

In the report, the GAO outlined eight priority recommendations for the Office of Science and Technology Policy (OSTP), focusing on managing climate change risks, utilizing data to meet national goals, addressing research security risks, and regulating AI. The GAO revisited a December 2023 recommendation advising OSTP to inform agencies of their status as implementing agencies with regulatory authority over AI.

OSTP disagreed with this recommendation, citing a Trump-era executive order on maintaining leadership. While OSTP and a select committee on AI must identify which agencies are implementing agencies, they are not obligated to communicate this designation to those selected.

Some agencies expressed that they were unaware of their designation, leading to uncertainty about their requirements under the memorandum. The report emphasized the need for full implementation of this recommendation to ensure agencies can fulfill their obligations related to the executive order.

GAO stated that OSTP’s action to inform agencies would enhance their efforts to proactively prepare for the complexities, benefits, and risks associated with AI.

In response to the GAO, OSTP highlighted that implementing agencies must fulfill six responsibilities, including promoting sustained investment in AI research and development collaboration, a concern for OSTP. In a previous interview, OSTP Director Arati Prabhakar noted that much of the office’s work on AI through R&D occurs at the intersection of AI, information technology, and real-world applications.

“To do this work well is going to require all the marvels of AI,” she stated. “And it’s going to continue to require the laboratory research that’s been the foundation of so many of these different disciplines.”

• 0 Comments
• Ai Process
• Artificial Intelligence