OMB Issuing Updated Guidance on AI Use Cases

The Trump Administration issued updated guidance on artificial intelligence (AI) and how agencies should be reporting use cases of the burgeoning technology. 

The guidance from the Office of Management and Budget was shared internally but has yet to officially be made public. FedScoop obtained access and reported on the details. 

According to FedScoop the guidance is similar to that issued by the Biden Administration, but is slimmed down a bit.

Agencies will be required to provide much of the same information on AI that was required under the Biden Administration, including stage of development, whether it was developed in-house or purchased, and whether the use case involves personally identifiable information maintained by the agency, among other categories.

Slimmed Down Version

There are changes however. 

Instead of designating specific risk management practices for use cases that are rights and safety impacting, the Trump Administration is designating such uses as “high-impact. For such cases, agencies will be required to answer additional questions on the use case to ensure that it protects from harm. 

In addition, the Trump Administration moved to streamline reporting, replacing individual questions about dates of deployment in some areas, with a single question about operational development. Also removed is a question about what Procurement Instrument Identifiers were used. Instead, agencies have a prompt to disclose the vendor. 

And other categories were removed completely, including whether information sent out to the public followed federal quality guidelines and whether agencies had access to request computing resources for training and development. 

Senate Rebuffs AI Regulation Ban

Meanwhile, the Senate voted 99-1 to ban the Trump Administration from imposing a ten year moratorium on AI regulation as part of the amendments voted on during the “Big Beautiful Bill.”

The proposed ban would have barred states from passing their own AI-related laws for a decade, meaning no action could be taken on a wide variety of issues including political misinformation and explicit deepfake content.

AI TRAINING, WORKSHOPS OFFERED

And organizations continue to make offerings for public employees grappling with new AI technology and skills.

Innovate(US) provides no-cost, at-your-own-pace, and live learning on data, digital, innovation, and AI skills for public service professionals. The nonprofit provides a AI certificate, courses, live webinars, and recently announced new modules, Responsible AI for Public Sector Legal Professionals.

Previous
Previous

Federal Workforce Data Website FedScope to Get a Redesign Amid Workforce Upheaval

Next
Next

Bill Would Force High-Ranking Feds to Take Annual Training on Budget Process