In our 'Contentious AI' insight series we unveil key facts about the development, adoption and use of AI and how it can become contentious.
In this article, we explain “AI washing” and how to mitigate it.
Artificial intelligence (AI) has rapidly transitioned from an emerging concept to a common feature of business operations. Many organisations now deploy AI systems across products, services and internal processes, and increasingly refer to those capabilities in public disclosures, marketing materials and investor communications.
However, where those references overstate or mischaracterise the role AI actually plays, companies may be exposed to a growing category of legal risk. This conduct is known as “AI washing” and it is attracting increased scrutiny from regulators and plaintiff law firms.
The scale of adoption and AI washing
In the last 12-18 months, three conditions have converged to create serious legal exposure for Australian businesses.
First, AI adoption in Australia is near-universal. Companies are embracing AI to improve efficiency, cut costs, and gain a competitive edge. According to the Department of Industry, Science and Resources' 'AI Adoption Tracker', 82% of mid-sized Australian businesses (200–500 employees) have already adopted some form of AI.
Second, the commercial pressure to overstate is acute. A 2026 survey of 500 Australian start-up executives by investor management software company Carta found that 80% of participants believe there is an AI bubble and feel pressure from investors or the market to integrate AI into their operations.
Third, many boards and management teams do not fully understand how their AI systems operate? This is what is known as the black box problem. Machine learning models ingest massive amounts of data and identify complex patterns across vast datasets, but they do not show their work in a clear or detailed way. Accordingly, explaining why the algorithm reached a specific conclusion can be difficult. It is particularly problematic where companies are making public claims about AI systems they might not fully understand.
The upshot is that the pool of companies exposed to AI washing is potentially significant, and the legal framework to act on that exposure is already in place.
Overseas experience
Regulators and courts overseas are already reacting to AI washing, underscoring the need for Australian companies to be vigilant.
In March 2024, the US Securities and Exchange Commission (SEC) settled charges against two investment advisory firms, Delphia (USA) Inc. and Global Predictions Inc., for allegedly misleading investors about their use of AI. Delphia claimed its AI could analyse collective data to “predict which companies and trends are about to make it big” when it had never actually built the AI capabilities it described. Global Predictions falsely claimed to be the “first regulated AI financial advisor” and promoted “expert AI-driven forecasts” it could not substantiate. While the statements could be seen as being generic marketing statements, the SEC alleged that the companies misrepresented the use of their AI capabilities and effectively engaged in AI washing. Both firms were ordered to cease and desist and pay civil penalties.
In September 2024, the US Federal Trade Commission (FTC) announced a crackdown on deceptive AI marketing. As part of this crackdown, it commenced an action against DoNotPay, a company that claimed to offer an AI service that was "the world's first robot lawyer". According to the FTC complaint, DoNotPay promised that its service would allow customers to "sue for assault without a lawyer" and "generate perfectly valid legal documents in no time". However, the FTC asserted that the AI output could not deliver on these promises and was not a substitute for the expertise of a human lawyer. In February 2025, the FTC announced an order requiring DoNotPay to stop making such claims.
The legal framework in Australia and risks to companies
Australian regulators are following the same trajectory. Although there is no AI-specific legislation, consumer protection laws and directors' duties laws provide the tools to act, and Australian Securities & Investment Commission (ASIC), Department of the Treasury and the Australian Competition and Consumer Commission (ACCC) have each signalled that AI washing is firmly in their sights.
In August 2025, ASIC published a media release (25-171MR) noting that AI washing is one of the top five online investment scam trends it is targeting. ASIC specifically advised consumers not to believe "AI generated" returns or claims that an investment is "safe", noting that investment scams remain the leading scam type by losses in Australia, with Australians losing $945 million to investment scams in 2024.
In October 2025, the Department of the Treasury released its 'Review of AI and the Australian Consumer Law (ACL)', which highlighted that Australians enjoy strong protections for AI-driven products and services under the ACL, just as they do for traditional products. In essence, companies (and their directors) that make false or unfounded claims about AI capabilities face the same legal exposure as they would for any other misleading statement. This means directors and companies who permit misleading statements about AI capabilities may find themselves exposed under both the ACL as well as other legislative provisions, including the Corporations Act 2001 (Cth).
Significantly, the Treasury also specifically identified the opacity and lack of explainability in decisions made by advanced AI models (the ‘black box problem’) as a key risk regulators are seeking to address. For companies that cannot explain how their AI systems reach decisions, that regulatory focus compounds the exposure – not only may the public-facing claims about those systems be tested, but the inability to account for the underlying decision-making will make claims harder to defend.
In December 2025, the ACCC dedicated a full section to AI washing in its industry snapshot and named it as a key consumer risk. The ACCC noted that firms are incentivised to overstate AI functionality because consumers pay more for products purporting to have advanced AI capabilities than for comparable products that do not. The ACCC also flagged that companies are inaccurately promoting AI systems as being autonomous and requiring limited human supervision – also known as "agentic" AI – in order to capitalise on the recent hype. The DoNotPay case was cited as illustrative of the conduct the regulator has in its sights.
Beyond regulators, plaintiff law firms (and it might be assumed also litigation funders) are keeping a close eye on AI washing. In a recent Sydney Morning Herald article – "Class action lawyers set sights on Atlassian" (31 March 2026) – it noted that class action firms are probing software companies (following steep recent sell-offs across the sector) to determine whether they have overstated the capabilities of their own AI tools and engaged in AI washing.
If a regulator penalises a company for misleading AI claims, a risk exists that it is followed by “piggyback” class actions filed on behalf of consumers or investors echoing the regulator’s findings. Shareholder class action risk also exists for listed entities. If a company makes representations to the market about the quality or functions of the AI tools, practices or capabilities it is implementing, and the purported benefits (financial or otherwise) associated with their implementation, this may increase investor sentiment. However, if those representations are later found to be inaccurate and/or misleading, and upon the market finding out the truth the share price of that company declines, this presents clear class action risk.
Practical steps to mitigate exposure
We suggest companies take the following steps:
- Before you announce or market anything related to AI, ensure it accurately describes what your AI does (and potentially does not do) using evidence you can rely upon if tested.
- Be conscious to avoid vague or overgeneralised assertions about your AI systems, and to not omit important qualifiers to your AI tools. Disclaimers can be useful.
- Establish a regular review cycle for any descriptions of AI in your products, services, marketing or investor disclosures - AI technology is evolving quickly and a statement about your AI that was true a year ago may no longer be true today.
- Make sure your team maintains a strong understanding of how your AI systems work and make decisions. This not only helps you avoid inaccurate claims, but will also put you in a much better position to defend your AI’s use if questions arise later.
- Establish board-level governance criteria regarding AI, including regular, structured briefings on the AI systems the company uses, the claims being made about them publicly, and any material changes to their performance or scope.
- When conducting an internal audit of your AI capabilities or reviewing past public statements, consider involving in-house counsel and/or engaging external legal counsel to direct the process. Reviews conducted for the dominant purpose of obtaining legal advice may attract legal professional privilege, protecting documents produced as a result of the audit from production in subsequent litigation or regulatory proceedings.
Coming up next
The insight series will expand to explore the following topics:
Navigate your AI pathway with confidence and clarity.
Contact us before AI becomes contentious in your organisation.
