Seeing Like a Dashboard: Digitalized Due Diligence, Legibility, and the Limits of Automated Human Rights Assurance in the EU's Twin Transition

The EU frames digitalization as the way to make sustainability due diligence affordable, and as the Omnibus I package narrows the Corporate Sustainability Due Diligence Directive (CSDDD) and recasts it around simplification, automation is offered as a way to do more with less. Drawing on the experience of building a sixteen-agent system for automated supply-chain due-diligence verification, this position paper argues that digitalizing human rights and environmental due diligence (HREDD) opens a structural gap between what such systems can see and what due diligence exists to find, and that the gap does not fall evenly across the value chain. The system reasons over three machine-legible input classes, supplier self-reports, structured ESG disclosures, and spend-based proxies, which by construction exclude the lower tiers and informal actors where the gravest harms concentrate. We give a legibility-gap characterization mapping each targeted harm to whether it produces machine-readable signal, and identify three side-effects that emerge only in implementation: amplified self-reporting bias, diffused responsibility, and a perverse interaction with simplification that shifts the burden of proof onto frequently Global South suppliers while calibrating the tools of proof to metropolitan data conditions. We argue that what a system is built to measure is a governance choice, and that automation must be an input to, not a substitute for, human rights due diligence.