Case 2: The Algorithmic Boss — Innovare AI Analysis

As you read — hold this question

If a system can measure everything a worker does, what does it still fail to measure — and what happens to the people managed by the gap?

fired by app

Stephen Normandin drove for Amazon Flex for four years. Then an automated system decided his numbers were too low and terminated him — by email. No manager. No appeal that a human read.

Performance management is AI's second HR beachhead. The promise is objective, real-time measurement. The risk is that the measure becomes the manager — and the human disappears from a consequential decision. Normandin appealed for months and eventually emailed Jeff Bezos (Bloomberg, 2021).

The course view

The textbook versionQuiz: augmentation vs substitution

AI in performance management & engagement

The course frames AI performance tools as decision-support: continuous feedback, skill-gap detection, predictive retention. Engagement tools promise to sense morale and personalise development. Done well, this is augmentation — freeing managers from admin to coach people.

The examinable point: AI here is meant to inform the manager, not replace the managerial judgment that weighs context, fairness and circumstance.

Carry this

Performance and engagement are areas 2 and 3. The course's own standard is augmentation — AI informs the human decision, it does not become it.

The idea being revived

The rest of the story

Digital Taylorism — a 1911 idea with 2020s sensors

Frederick Taylor's Scientific Management (1911) broke work into measured, timed units to be optimised from above. Algorithmic management is digital Taylorism: the same logic, now with location pings, keystroke logs and productivity scores. In 2020 Microsoft shipped a Productivity Score that could rank individuals; after a public backlash it rolled back the individual-level metrics (GeekWire, 2020).

The proxy problem

Systems measure what is easy to count — deliveries per hour, messages sent — not what matters (judgment, care, teamwork). When a countable proxy becomes the target, people optimise the proxy and the real goal drifts. That is Goodhart's Law wearing a productivity dashboard.

What the evidence and the humane view say

The other side

People resist the robot boss — and the fix is the Progress Principle

Cornell research (Schlund & Zitek, Communications Psychology, 2024) found people resist decisions made by algorithms far more than identical decisions made by humans — roughly 30% vs 7% in one measure — because they feel reduced to a number.

The constructive counter is Harvard's Progress Principle (Amabile & Kramer, HBR, 2011): the single biggest driver of good inner work life is making progress on meaningful work. If you genuinely cared about a worker, you would measure and remove the friction in their way — not just count their outputs. The humane use of AI here is to surface blockers, not to tighten the whip.

The other side

Almost any proxy for wellbeing can introduce bias. The test of an algorithmic boss is simple: does it help people make progress, or just prove they are being watched?

Quick recall — without looking back

Test yourself on this case

Question 1 of 3

What is “digital Taylorism” and which company had to roll back an individual productivity metric?

Algorithmic management applying Taylor's 1911 measure-and-optimise logic via modern sensors. Microsoft rolled back the individual-level Productivity Score after a 2020 backlash.

Question 2 of 3

What did the Cornell (Schlund & Zitek, 2024) research find about algorithmic managers?

People resist algorithmic decisions far more than identical human ones (~30% vs 7% in one measure) — they dislike being reduced to a number.

Question 3 of 3

What is the Progress Principle, and why is it the humane counter here?

Amabile & Kramer (HBR 2011): the biggest driver of good inner work life is making progress on meaningful work. The caring use of AI is to detect and remove blockers, not merely to count outputs.

Module 9

← Case 1: The Hiring Machine

Case 3: The Engagement Illusion →

Module 9 Videos

Module 9 · Short · The Algorithmic Boss

Module 9 · Long Form · The Algorithmic Boss (Short-Doc)

Sources

Module content: BUSN9049 Module 9 — performance management & engagement (deck + videos). Flinders University, 2026.
Algorithmic management: Bloomberg, “Fired by Bot” (Stephen Normandin, 2021); GeekWire, Microsoft Productivity Score rollback (2020); Taylor, The Principles of Scientific Management (1911).
Evidence: Schlund & Zitek, Communications Psychology (Cornell, 2024); Amabile & Kramer, “The Power of Small Wins,” HBR (2011).