The Residual Human Burden Index (RHBI)

Overview Artificial intelligence is marketed and measured as a technology that removes human work. Vendors, enterprises and public institutions justify deployment through gains that are visible and immediate: output produced, time saved, headcount reduced. This paper argues that those metrics capture only one side of an accounting identity. Across knowledge work, automation frequently does not remove human work; it relocates it into less visible forms, namely verifying outputs, correcting errors, proving and defending decisions, absorbing liability, and carrying the cognitive and emotional load of supervising a system that cannot be fully trusted. The Residual Human Burden Index (RHBI) We introduce the Residual Human Burden Index (RHBI), a composite measurement framework that quantifies the human work remaining after an artificial intelligence system is introduced. RHBI decomposes this residue into six measurable classes: verification burden, correction burden, liability burden, cognitive burden, emotional burden and institutional repair burden. Each is scored from 0 to 100 and combined through a weakest-link cascade, because a single severe class, most often liability, can negate the value of an otherwise efficient deployment. Burden Displacement Mechanism The framework’s theoretical core is the burden displacement mechanism. Gross productivity gains are immediate, attributable and dashboard-visible, while residual burdens are lagging, diffuse and borne by individuals downstream. A deployment can therefore report success while accumulating stress, legal exposure and operational fragility beneath it. RHBI pairs the burden score with a net automation value test, converting the central question from whether an artificial intelligence system performs a task to who pays the human cost after the machine acts. It is proposed as a missing governance and procurement layer for vendors, enterprises, auditors and regulators. Keywords: Residual Human Burden Index; AI Automation Burden; Burden Displacement; False Productivity; Human Verification Cost; AI Liability Transfer; Human Agency; AI Governance; Hidden AI Labor; Net Automation Value

Hali tarjima qilinmagan