Everybody wants artificial intelligence to save the economy. Governments bank on it, tech firms pitch it as the cure-all for stagnation, and investors throw billions at datacentres. But reserve bank heads are starting to look at the numbers and realize nobody is actually getting more work done.
Reserve Bank of Australia governor Michele Bullock recently laid out a blunt truth that official budget projections prefer to ignore. Artificial intelligence is not driving productivity gains right now. In fact, it might be an overhyped bubble adding to inflationary pressures instead of solving them. When central bankers start questioning the shiny new toy, it's time to pay attention.
The Productivity Illusion That Everyone Keeps Buying
Governments love long-term economic models because they let politicians project prosperity decades into the future without paying bills today. Australia's intergenerational projections rely heavily on the assumption that artificial intelligence will trigger a massive surge in economic activity, lifting per-person output dramatically by 2066.
There is just one problem. The data doesn't back it up.
Bullock pointed to research from South Korea showing that employees adopting artificial intelligence tools often produce the exact same output while spending roughly an hour and a half less per week actually completing tasks. Workers fiddle around with prompts, test automated workflows, and experiment with code generators. Output doesn't jump. It flatlines or dips.
Companies buy expensive subscriptions, upgrade cloud infrastructure, and build power-hungry datacentres. That capital expenditure pushes up inflation before it yields any tangible efficiency gains. You spend more money to get the same output. That is the exact opposite of productivity.
Bubbles Built on Hope and Hype
Markets thrive on narrative. Meta debuts a new conversational agent, stock prices jump double digits, and Wall Street acts like the future has arrived. Central banks globally are watching this frenzy with intense anxiety.
Bullock noted that central banks view artificial intelligence as the great white hope for modern economies. Everyone desperately wants it to work because traditional economic levers are exhausted. But hope is not an asset class.
If tech valuations correct sharply, the fallout won't stay confined to Silicon Valley. Broad market slumps drag down consumer confidence, tighten lending conditions, and squeeze businesses that overinvested in unproven software stacks.
What Founders and Leaders Get Wrong About Adoption
Most companies approach artificial intelligence backwards. They buy enterprise licenses, hand them to employees, and tell them to find ways to be innovative. That strategy fails every single time.
Tools don't fix broken workflows. If your internal communication is a mess, your project management is unstructured, and your documentation is outdated, throwing smart assistants at the problem just accelerates chaos.
Real efficiency requires a complete redesign of how businesses operate. You have to map out bottlenecks, rewrite operational procedures, and train teams on strict integration protocols. Until organizations do that hard internal plumbing work, automated tools remain expensive toys rather than profit drivers.
Housing, Immigration, and Real Economic Pressures
The disconnect between official projections and ground reality extends beyond technology. While politicians debate immigration targets, the strain on physical infrastructure tells a different story.
Immigration has driven the bulk of Australia's economic growth over recent years. New arrivals work and spend, supporting businesses across the board. But housing supply cannot keep pace. When population growth outstrips residential construction, the cost of living spikes, putting immense pressure on household budgets.
Property markets are already feeling the shockwaves. With interest rates sitting at multi-year highs and investor tax break reforms kicking in, home loan approvals have dropped sharply. Bullock pointed out that while the housing downturn falls within historical parameters, the affordability crisis remains severe.
Where We Go From Here
Stop treating technology as a magic wand. If you run a business, audit your operational processes before writing another check for software licenses. Measure actual output per hour, not just tool adoption rates.
Ignore the marketing decks promising instant transformation. Real productivity gains require boring, systematic restructuring of workflows. Until companies do that work, central banks will keep warning us about bubbles, and the productivity miracle will remain just out of reach.
