The Biden administration is calling on agencies to move on from just talking about artificial intelligence and actually use these tools for their day-to-day work.

The White House in October released its blueprint for a “Bill of Rights”  for the design, development and deployment of AI and other automated systems.

The AI Bill of Rights outlines what more than a dozen agencies will do to ensure AI tools deployed in and out of government align with privacy rights and civil liberties.

The administration, as part of those efforts, is working on new federal procurement policy and guidance, to ensure agencies purchase and implement AI and automation tools that transparent and free of bias.

Sorelle Friedler, the assistant director for data and democracy at the White House Office of Science and Technology Policy, said Monday that the blueprint is “putting the weight of the White House” behind a policy area that’s provoked a lot of conversation, but hasn’t led to widespread implementation across government.

“We are not really breaking new ground, but adding to the conversation and helping us move the conversation forward, from principles into practice,” Friedler said at a Brookings Institution event unpacking the AI Bill of Rights.

The Biden administration also released a technical companion to the blueprint that serves as a roadmap for implementing transparent and accountable AI tools in government.

“We are also trying to live up to that across the federal government,” Friedler said.

Friedler said the non-binding policy document is focused on putting a more governmentwide focus on the automated systems that have the “potential to meaningfully impact individuals or communities’ rights, opportunities or access” to government benefits and services.

“More and more, we’re seeing these technologies drive real harms. Harms that run counter to our core democratic values, include the fundamental right to privacy, freedom from discrimination and our basic dignity,” Friedler said.

A panel of AI experts said the blueprint for an AI Bill of Rights makes technology a focus of the administration’s ongoing efforts to ensure underserved communities have equal access to federal benefits and services.

President Joe Biden signed an executive order on his first day in office mandating agency reports on ways to ensure “equitable delivery of government benefits.”

But Harlan Yu, the executive director of Upturn, said many of those agency reports released this year didn’t address how AI and related technologies can perpetuate some of those barriers to access.

“You could probably count on one hand — a few fingers, really — how many of those plans address technology issued at all,” Yu said. “What that told me was that the federal agencies aren’t being attentive enough to these issues, that we really needed leadership at the highest level of government coming from the White House to draw political attention to these issues, and to make it a priority across the federal government.”

Yu said Biden’s executive order also underscored how many federal agencies don’t have the expertise to address many of the technological and ethical barriers that prevent them from adopting AI tools.

“That’s something that we need to fix, longer term,” he said.

The AI Bill of Rights marks the Biden administration’s first major effort to address the use of this technology across agencies. But the Trump administration issued two executive orders on the topic.

Alex Engler, a senior governance studies fellow at Brookings, said agencies didn’t prioritize implementation of the Trump administration executive order requiring them to document their AI uses cases.

“They kind of complied, but [it was] really the absolute bare minimum,” Engler said.

AI.gov currently lists about 100 AI use cases across federal agencies.

Engler said the blueprint for an AI bill of rights comes at an “absolutely necessary time” to unpack the ethical considerations of how agencies should use AI tools.

“A lot of people might be a little exhausted with AI ethics and AI principles. That is common, if you work in this space. But it is fundamentally different that the federal government did it so thoroughly, did it so carefully, in a way that will drive not only governance, but also hopefully, some independent use of algorithms,” Engler said.

Jerome Greco, a digital forensics supervising attorney with the Legal Aid Society, said the White House putting out this non-binding policy document helps bring government regulation into focus around this technology.

“It means that there’s effort, money and attention being put into this. It also means that other people are going to pay attention to it. Whether or not that will cause a long-lasting impact remains to be seen, but it has the ability to do that. It forces people to acknowledge what is being said there, and either embrace it or have to combat it,” Greco said.

While the AI Bill of Rights outlines AI ethical considerations for much of the civilian federal government, the document stops short of covering federal law enforcement and national security activities.

“Obviously, in those areas, that’s where a lot of these protections are most necessary,” Yu said.

The intelligence community and Defense Department released their own AI policies last year. However, Greco said there are currently few legal guardrails on how federal law enforcement agencies use facial recognition algorithms as part of their investigations.

“The [AI] Bill of Rights is not binding. This is not legislation. It could lead to that, and I’m hopeful that it does on many fronts, but it currently doesn’t. And our courts are not set up to handle these things,” he said.

“By exempting them, I think it emboldens them, because they’re just so used to having everything be exempted from, to not being held accountable, not to have any sort of transparency. And by making that explicit exemption really just bolstered that,” Greco added.

Friedler said the administration is continuing to meet with companies on developing their own ethical AI principles.

“There are a lot of chances for innovation and making this an industry competitive advantage, to really make AI that people trust, that respects our rights,” Friedler said.

Yu said the blueprint represents “mile one of a long marathon,” and the impetus for the administration to introduce rulemaking and guidance, and for Congress to introduce legislation.

“This document, in the long term, will be judged not by what’s on paper, but all the concrete actions that are going to flow from this document, particularly from the federal agencies,” Yu said. 

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.