SUN, OCTOBER 18, 2026
Independent · In‑Depth · Practitioner‑Tested
✎ News

Amazon Built a Model That Scores Warehouses by Union Risk

An internal Amazon system called Atlas scores warehouse sites between 100 and 900 according to the likelihood of union organising, with two UK sites flagged at the highest tier, according to documents reported by openDemocracy.

By AIToolsRecap October 18, 2026 5 min read 24 views
Home › Articles › News › Amazon Used AI to Score UK Warehouses by Union ...

Amazon operates an internal system called Atlas that scores warehouse sites on the likelihood of union organising, according to documents reported by openDemocracy. Sites are scored on a scale from 100 to 900, and two UK sites are flagged at Tier 1, the highest risk band.

What is actually new here

Employers have monitored organising for as long as there has been organising, so the activity is not novel. What a scoring system changes is timing and scale.

A number attached to every site, updated continuously, lets a company act before anything visible happens — redirect management attention, change scheduling, move staff — at a point where there is no organising campaign to respond to and therefore nothing for a regulator to examine. The conduct that labour law governs is mostly conduct during a campaign.

Joining a union is a legally protected activity in the UK. A risk score turns that protected activity into an operational metric, which is a materially different thing from a manager noticing discontent.

The wider pattern

This sits with the week's other workplace-AI stories rather than apart from them: Anthropic publishing that agent hours run six times employee hours, and its junior hiring share falling from 27% to 9%. Different companies, different mechanisms, same direction — the employment relationship is increasingly mediated by systems the employee cannot see and cannot contest.

Amazon has not publicly commented on the Atlas documents at the time of writing.

Sources: openDemocracy

Tags
AI News2026AI agents
⚑

Spot an inaccuracy?

We verify facts before publishing and correct errors promptly. If something in this article is wrong or outdated, let us know.

Report an error →