AI slop responses to RFPs, RFIs are overwhelming federal agencies
Inundated with AI-generated responses to industry outreach efforts, agencies are considering how to best address the growing challenge.
Agencies use requests for information as a market research tool to explore possible solutions and to see what’s available in the market. Requests for proposals are another lever agencies lean on in the acquisition process to assess options and evaluate approaches.
But the rise of generative AI has complicated the efforts, according to Anil Chaudhry, senior advisor for AI, autonomy and national security for the Department of Transportation.
“We used to get maybe 10 responses to an RFI or RFP,” Chaudhry said Wednesday onstage during the SAP Public Sector Summit in Washington, D.C. “Now, we’re getting 500 because everybody puts it in a chatbot and sends it to us.”
Agencies are finding it challenging to sift through it all.
“Everybody in government has these issues where they’re getting overwhelmed by qualitative information,” Chaudhry said. “Trying to deal with that volume, it’s a whole-of-government problem, and we have whole-of-government teams working on these problems.”
Private-sector companies are overflowing with generative AI tools that produce lengthy, oftentimes glib chunks of texts. Vendors have even begun catering to companies that want to speed up the response process.
InspectMind, for example, has a new product that is marketed as a system that can easily create first drafts for RFI responses. Nomic touts an AI agent that drafts RFI responses, reducing time spent on the task. Datagrid is another AI vendor looking to capitalize on the growing market.
While industry has created the influx, it could also aid agencies in scrutinizing the large volume of content.
In a recent RFI, Immigration and Customs Enforcement asked interested vendors to help the Department of Homeland Security component identify methods that would distinguish qualified suitors from AI slop.
“The Government recognizes that the proliferation of generative AI tools has made it easier to produce highly polished, technically accurate written proposals,” ICE said in its request to industry. “To ensure the Government selects a vendor with genuine, proven capabilities to execute the scope of work, what specific evaluation methods… do you recommend to differentiate genuine, demonstrated hands-on capability features creation from purely generative AI responses?”
Combatting AI with more AI is on the table, too.
Earlier this year, a Defense Department request for proposals included a disclosure of potential AI use in the review process.
“The Government may employ AI as a tool to assist in the analysis and review of offeror proposals,” DOD said in accompanying documents, pointing to a variety of use cases, such as summarizing proposal content, identifying compliance with requirements, and highlighting areas of potential strength, weakness, or risk.
“Notwithstanding the use of AI tools to support the evaluation process, all final evaluation judgements and source selection decisions will be made exclusively by duly appointed Government personnel,” the agency added.
The strategy has been met with some skepticism from lawyers and stakeholders.
“It is not difficult to imagine the potential for generative AI tools to hallucinate or otherwise mischaracterize proposal details that are material to an agency’s award decision,” Michael Brewer, a member of Burr and Forman’s construction practice group, said in an April blog post.
