Australian competition minister warns AI tollbooths risk widening wealth gap

Federal Assistant Minister for Competition Andrew Leigh has warned that concentrated control of artificial intelligence could expand inequality between capital owners and workers, urging policymakers to examine market structures before tech platforms dominate the broader economy.

Speaking ahead of an address to an investment conference on Tuesday, Leigh argued that while machine learning can help junior employees narrow the skills divide with senior staff, the financial gains of automation risk flowing almost entirely to those who own the core technology. He warned that if everyday businesses depend on a tiny pool of upstream model builders, those platforms will begin operating like tollbooths across the economy.

The shift from wages to capital

Leigh noted that previous technological shifts created large fortunes for a small cohort of industrialists while workers competed over remaining earnings. Automation reduces the reliance of businesses on manual and routine human input, which means returns compound for those who hold equity in foundational systems.

While early Australian workforce data shows steady entry-level employment compared to softer youth hiring signals in the United States, the longer-term economic risk lies in vendor dependence. When software providers control the underlying intelligence behind retail checkouts, professional advice and back-office tools, they hold pricing power over every firm relying on their infrastructure.

What this means for Australian businesses

For Australian small and medium enterprise owners, this regulatory focus points to practical steps to keep control over company operations and budgets:

• Avoid relying on a single proprietary AI vendor for critical business processes, as vendor lock-in leaves your operations exposed to price hikes and sudden feature changes.
• Keep business workflows modular by using open frameworks or software that can swap underlying language models with minimal code rewriting.
• Focus internal automation efforts on building unique business data assets rather than standard prompts that any competitor can rent from the same provider.
• Review software contracts carefully to ensure your operational data remains your property and cannot be used to train models that benefit competing platforms.