The Pamir Intelligence Platform, a SaaS database that uses machine learning to surface forced-labor signals on Chinese-language social media, is now in use by the U.S. Department of Homeland Security's UFLPA Entity List Office, founder Daniel Murphy said Tuesday. The goal, he said, is to turn "fragmented evidence" into "clear, entity-level intelligence" that helps the Entity List Office identify companies implicated in the use or facilitation of forced labor.

Murphy developed the methodology in 2023 while investigating labor conditions in Chinese seafood processing plants. The platform scrapes, parses and categorizes unstructured video, image and social-media data from Chinese platforms — content he says other risk-intelligence vendors have "completely under-leveraged."

What does Pamir offer that existing traceability tools miss?

Murphy was categorical: Pamir is not a traceability platform.

"We don't map supply chains," he said. "We'll never have one of our customers come onto our platform, conduct a risk assessment, and be told there may be this entity in their supply chain. We only risk assess based on the entities that we're provided visibility on; we're not extrapolating your Tiers 4, 5, 6."

What Pamir does offer, per Murphy, is site-level intelligence:

• Identification of who is working at a given site
• Where those workers came from
• Recruitment and transfer mechanisms
• Government bodies or intermediaries involved

The platform is "human first," he added. Every piece of scraped material passes through human analysts because "there can't be any 'AI slop in the picture'" when findings trigger C-suite decisions.

How does this fit into DHS's wider enforcement stack?

The Entity List Office sits under DHS's interagency Forced Labor Enforcement Task Force. In July, the task force added 43 entries to the roster — the largest expansion since the list's creation and its first update since the end of the Biden administration in January 2025.

Customs and Border Protection has previously contracted:

• Oritain for origin-testing services
• Exiger, a supply-chain risk-management company whose AI flags illicit transshipment

Pamir layers on worker-movement visibility. "Shipping documents don't show where workers go," Murphy said.

Human-rights groups have flagged uneven Entity List enforcement as a problem: companies can route forced-labor-tainted materials through unlisted intermediaries in third countries, exploiting the difficulty of mapping complex multi-country chains. Murphy's approach complements physical-goods tracing by tracking the workforce rather than the freight.

Is UFLPA enforcement shifting beyond Xinjiang?

The July list update included entities operating outside the Xinjiang Uyghur Autonomous Region. Murphy reads that move as confirmation of a pivot DHS signalled through border actions.

"DHS has signaled in both its enforcement strategy and its actual actions in terms of Entity List decisions and seizures at the border that the labor transfer system is as much of a concern as the forced labor practices within the region," he said.

Beijing frames its labor-transfer programs as voluntary poverty alleviation. Murphy argues that relying on Chinese state media alone lets authorities curate that picture; videos from Uyghur and other minority workers on their own feeds are the harder signal.

What should sourcing and compliance teams take from this?

Risk is geographically concentrated rather than uniform, Murphy said. When Pamir maps the labor-transfer system at sub-prefecture level, the platform surfaces risk in specific districts. "You see that risk is very highly concentrated in specific districts of China," he said. "That makes it almost predictable, or at least you're able to model risk based on the location of a supplier."

He expects more Entity List additions. The sourcing and compliance decision the news forces is straightforward: brands with Chinese exposure should add workforce-movement diligence to existing material-traceability workflows and anchor that diligence to the supplier's sub-prefecture district rather than the country of origin alone.