The ACCC Just Made AI-Washing a $100 Million Problem

On 28 March 2026, the Treasury Laws Amendment (Doubling Penalties for ACCC Enforcement) Act doubled the maximum corporate penalty under the Australian Consumer Law from $50 million to $100 million per contravention. For the most serious breaches, the penalty is the greater of $100 million, three times the benefit obtained, or 30 per cent of adjusted turnover. Individual penalties remain at $2.5 million per breach.

In February 2026, the ACCC published its own AI transparency statement and flagged AI-washing — misleading claims about AI capabilities — as an enforcement concern. The regulator's 2026–27 compliance and enforcement priorities, which took effect 1 July, specifically target manipulative conduct in digital markets: dark patterns, subscription traps, and misleading pricing. As ACCC Chair Gina Cass-Gottlieb put it: "Competition alone is not enough. Markets only deliver when people trust them."

Translation: if your website says "AI-powered" and the underlying technology is a rules-based spreadsheet macro, you now face the same penalty framework that produced a $125 million fine against Volkswagen and a $60 million fine against Google.

AI-washing is not a fringe problem. A 2026 Carta survey of 500 Australian startup executives found that 80 per cent believe there is an AI bubble, yet feel pressure from investors or the market to integrate AI into their operations. More than 40 per cent said AI has already changed how they position their business when raising capital. The incentive to overstate is structural.

Common examples that Australian regulators are watching: advertising products as "AI-powered" when the underlying functionality is standard automation or rules-based workflows. Claiming full automation when the system actually relies on extensive human intervention — the Amazon "Just Walk Out" case, where 1,000-plus workers manually reviewed roughly 70 per cent of transactions, is the global reference point. Promoting tools as "agentic AI" to capitalise on recent hype when they are basic task-completion scripts. Making performance claims — "our AI reduces errors by 40 per cent" — without verified data to support them.

For an SME, this is not theoretical. If your proposal documents claim AI-driven scheduling, AI-powered quoting, or AI-enhanced compliance monitoring, the question is whether those claims describe what your systems actually do. The same legal standard that applies to a pharmaceutical company claiming clinical efficacy applies to a trades business claiming AI optimisation.

In October 2025, the ACCC commenced Federal Court proceedings against Microsoft for allegedly misleading 2.7 million Australian subscribers when it bundled Copilot into Microsoft 365 plans. The allegation: Microsoft told Personal and Family subscribers they must accept Copilot integration at a 29 to 45 per cent price increase, or cancel — without disclosing a third option, the Classic plans, which retained existing features at the original price.

The Classic plans were accessible only deep within the cancellation workflow. Microsoft has since apologised and offered refunds. But the case is proceeding — and under the new penalty regime, the maximum exposure is dramatically higher than when the conduct occurred.

This matters because it establishes that AI-related consumer deception is being actively litigated in Australia, not just discussed. Class action firms are probing Australian software companies to determine whether they have overstated their AI capabilities. If a regulator penalises a company for misleading AI claims, a risk exists that piggyback class actions follow.

The same framework protects you as a buyer. If a software vendor tells you their product uses machine learning to optimise your scheduling and it turns out to be a series of if-then rules, that is potentially misleading conduct under the ACL. If a platform charges you a premium for "AI features" and those features are basic automation you could replicate in a spreadsheet, the representation is in scope.

We covered in a recent note how Microsoft and Google both bundled AI into standard subscriptions at 12 to 50 per cent price increases. The ACCC's case against Microsoft is not about the price increase itself — it is about the failure to disclose alternatives. For every vendor pitching you AI capabilities at a premium, the relevant question is now: what specifically does the AI do, how was the performance claim verified, and what is the non-AI alternative?

The Unfair Trading Practices Bill 2026, introduced to Parliament on 1 April, would further prohibit manipulative practices including subscription traps and dark patterns — with penalties of up to $50 million or 30 per cent of turnover. If passed, it is expected to commence 1 July 2027. The direction is unambiguous: consumer protection for digital services is tightening, not loosening.

First, audit every AI claim your business makes — website, proposals, pitch decks, social media. For each claim, ask: is this actually AI (a system that learns and adapts), or is it automation (a system that follows pre-programmed rules)? If you cannot clearly articulate the difference for a specific claim, rewrite it or remove it. The cost of overstating is now measured in hundreds of thousands at minimum.

Second, question your vendors. When a software provider tells you their product is AI-powered, ask three questions: what data does the model train on, what decisions does the AI make versus a human, and where can you see verified performance data? Vendors making legitimate AI claims will answer these without hesitation. Those who cannot are selling you a label, not a capability.

Third, document your methodology. If your business does use AI legitimately — and many trades and professional services firms now do, whether through platform features in Simpro, Xero, or Microsoft 365 — document what the AI does, what data it processes, and what performance improvements you have measured. This is not just legal protection. It is the basis for a credible competitive advantage that no competitor can undercut by simply adding "AI" to their website.

Key takeaways

Not sure whether your AI claims hold up?

Book a call and we will audit your current AI claims against what your systems actually do, identify where you are genuinely using AI versus automation, and help you build a positioning strategy that is both credible and legally defensible.