The Shadow Algorithm: How Government Agencies Are Using AI Without Transparency

The Department of Housing and Urban Development’s (HUD) use of artificial intelligence in policy decisions has come under scrutiny, with documents obtained through a Freedom of Information Act request revealing the agency’s reliance on AI tools. However, it is not just the potential biases or inaccuracies of these algorithms that are concerning, but also the lack of transparency surrounding their development and use.

HUD’s decision to rely on AI is part of a broader trend in government agencies embracing data-driven decisionmaking. But as seen in previous instances, the promise of AI can quickly turn to peril when its workings are shrouded in secrecy. HUD has withheld documents citing a “deliberative process privilege” that applies even to interactions with AI systems, raising serious questions about accountability and oversight.

The use of AI in policy decisions is not unique to HUD or a new development. Former members of the Department of Government Efficiency team at HUD demonstrated last year that AI can be a valuable tool for identifying agency rules ripe for rescission or revision. However, as staff attorney Tori Noble points out, “it’s not necessarily the case that we’d always know how tools are being used.” This lack of transparency makes the use of AI in policy decisions worrisome.

The government’s reliance on secrecy to justify withholding information about AI usage is particularly galling given the absence of clear guidelines or regulations governing its use. Currently, there are no laws requiring agencies to disclose whether AI has been used in creating rules, policies, or regulations. This lack of transparency creates a perfect storm of uncertainty and mistrust.

The use of AI in decisionmaking may be a fact of life, but it’s not too late for agencies like HUD to establish clear guidelines and protocols for its use. As lecturer Mark Fagan notes, “if AI is being used to assess policy as one of the tools in the toolkit, I think at this stage in the development and use of AI, it is good protocol to indicate that.” By establishing clear guidelines and protocols, agencies can build trust with citizens and lawmakers while ensuring that AI is being used as intended – to inform and improve policy decisions.

The issue at hand goes beyond simply whether or not AI should be used in policy decisions. It’s about accountability and transparency in government. If agencies are going to rely on AI tools, they must be willing to subject those tools to scrutiny and oversight. This means being transparent about how AI is developed, tested, and implemented, as well as providing clear explanations for its use.

As the government continues to push the boundaries of AI usage in policy decisions, it’s essential that we hold them accountable for their actions. The lack of transparency surrounding HUD’s use of AI raises serious questions about the administration’s commitment to openness and accountability. Agencies like HUD must prioritize transparency and establish clear guidelines for AI usage.

Secrecy will only serve to erode trust in government. By being transparent about their use of AI tools, agencies can build confidence in the decisionmaking process while ensuring that these powerful technologies are used as intended – to inform and improve policy decisions.

HUD and other agencies must now establish clear guidelines for AI usage. As we look to the future, one thing is certain: the use of AI in policy decisions will only continue to grow. But it’s not too late for agencies to get ahead of the curve and establish a culture of transparency and accountability.