Glean: AI creating hidden workload for Australian employees - IT Brief Australia

TEXT START: Australian knowledge workers are spending an average of 6.5 hours per week overseeing AI systems rather than using them to perform productive work, according to new research from Glean's Work AI Institute.

THE DISSECTION

The article documents a genuine transition cost: AI currently generates supervisory labour, error correction, governance overhead and low-value output. But it confines the problem to implementation quality. The proposed remedies—better tools, training, internal context and dynamic staffing—keep the corporation intact and position enterprise AI vendors as the technicians capable of repairing the disruption.

Its most revealing section is the discussion of AI analysing skills, relationships, availability and organisational participation to dynamically assemble teams. That is not merely productivity enhancement. It is the construction of a machine-readable labour market inside the firm: workers measured, ranked, recombined and made increasingly fungible.

THE CORE FALLACY

The article mistakes present human oversight for durable human economic necessity. “Bot sitting” is a lag phenomenon: the scaffolding required while systems remain unreliable. Under the Discontinuity Thesis, that supervisory layer is itself a target for automation once AI improves and competitive pressure makes human-only oversight too expensive.

The removal of “good friction” is treated as a skills-development problem rather than a participation-collapse mechanism. If AI lets firms obtain competent outputs without employing large numbers of junior workers, the formative labour market does not remain protected because it is educationally valuable. Cost competition will treat it as waste.

The article’s evidence does not show that AI has failed. It shows that deployment is incomplete and temporarily burdens humans before reducing their role.

HIDDEN ASSUMPTIONS

• Employers will recognise, compensate and redesign bot-sitting instead of using the workload as evidence that fewer workers are needed.
• Productivity gains will improve employee experience rather than accrue to owners through headcount reduction and tighter control.
• Human training and early-career work will retain market value even when AI can bypass much of the experience-building process.
• Better governance and training can solve the problem without confronting the competitive pressure toward deeper automation.
• Corporate hierarchy will remain fundamentally necessary, despite AI beginning to automate coordination itself.
• The vendor-sponsored survey is sufficiently transparent and independent to support broad conclusions; the supplied text provides no methodology capable of establishing that.

SOCIAL FUNCTION

Partial truth serving transition management, vendor-aligned propaganda and ideological anesthetic.

The article admits that AI creates hidden labour, burnout and degraded skill formation. That admission gives it credibility. It then relocates the structural threat into fixable categories—bad implementation, weak governance, poor training and obsolete org charts. The implied message is that the system can absorb the shock through better enterprise software.

That is commercially useful and strategically incomplete. It prepares institutions to manage the transition while avoiding the terminal question: what happens when the human oversight, junior training and coordination work being discussed are also automated?

THE VERDICT

The article is accurate about AI’s current tax and blind about its destination. Bot-sitting is not a durable new employment layer; it is the visible scaffolding of a system learning to remove humans from cognitive workflows. The same machinery that maps employees and dynamically staffs projects is building the control apparatus to measure, rank, substitute and eventually discard them.

“Good friction” will not survive a competitive mandate to minimise labour cost. This is not evidence that AI failed. It is evidence that the kill mechanism is still under construction.