If we treat current Australian AI equity efforts as over-focused on where AI runs (institutional licences, pilots, and platforms) rather than on who can reliably complete which tasks with AI, what new, task-completion–based equity metric (for example, “share of workers or students in each region who can use AI end‑to‑end for a short list of common work and coursework tasks”) would most usefully contradict existing state and federal deployment targets based on licences and pilots, and how might that contradiction change priorities for regional vs metropolitan policy and deployment design?

Answer

Define a simple, region-comparable metric:

"Task-completion equity index (TCEI)": for each region, the share of workers and students who can independently use AI end-to-end for a small standard bundle of tasks (e.g., draft–revise–finalise: a workplace email and summary; a basic report/lesson plan; a short coursework assignment), using any compliant access channel.

How it contradicts licence/pilot targets

• Licence/pilot metrics show where AI runs (accounts, pilots, hubs).
• TCEI shows who can actually finish standard tasks with AI.
• In metro hubs with many licences, TCEI will expose large internal gaps (high licence density, modest share of staff/students who can complete the bundle unaided).
• In some regional areas with shared or mediated access, TCEI will show higher-than-expected capability despite low licence counts.
• This creates a visible mismatch: states can "hit" licence/pilot targets while leaving most people unable to do basic AI-supported work/course tasks.

Resulting priority shifts

• From licences to people: targets move from "X licences/pilots in region" to "Y% of workers/students can complete the standard AI task bundle" in each region.
• Regional policy: emphasise shared access points, simple templates, and short coaching so more people can complete the bundle via libraries, TAFEs, councils and industry bodies, not just big metro institutions.
• Metro policy: push institutions to embed AI into core workflows and assessments (templates, expectations, guardrails) to raise the fraction of licence-holders who can actually complete the task bundle.
• Deployment design:
  • Set minimum TCEI floors by region and institution type.
  • Tie new public funding to increases in TCEI, especially in low-use regions and lower-status providers.
  • Require pilots to report gains in task-completion share, not only licence numbers or logins.