Algorithmic Accountability: Consumer Protection Law and AI in Australia
An online travel platform uses an algorithm that dynamically adjusts prices based on user behaviour: a customer viewing the same hotel ten times sees the price increase gradually, while a new visitor sees a lower price. The algorithm is designed to extract maximum willingness to pay from repeat customers. Is this unfair? Illegal? The Australian Competition and Consumer Commission (ACCC) says yes—and it is actively investigating these practices across the economy.
Algorithmic accountability has become the frontier of Australian consumer protection law. The ACCC, which regulates deceptive and unconscionable business conduct, is now explicitly treating algorithmic decision-making as subject to consumer law. If an AI system engages in misleading conduct, price discrimination that exploits vulnerability, or promotion of unsafe products, the business deploying the algorithm is liable—not the algorithm vendor, not the data scientist, but the business.
According to the ACCC’s 2024 Digital Markets Report, 71% of Australian businesses use algorithmic systems to inform consumer-facing decisions (pricing, targeting, recommendations, access). Yet only 31% can explain what their algorithms do, and just 18% have conducted algorithmic audits for fairness or compliance. This gap between widespread algorithmic use and minimal algorithmic governance creates urgent consumer protection risk.
ACCC Investigation Focus Areas
Algorithmic Pricing and Demand-Based Pricing
The ACCC is actively investigating algorithmic pricing—systems that set prices dynamically based on demand, inventory, user behaviour, or perceived willingness to pay. The legal question is: when is dynamic pricing unfair or unconscionable conduct under the Australian Consumer Law?
The ACCC’s position (issued in warnings and consultation papers, 2023-2024): dynamic pricing is lawful if prices are set transparently based on objective factors (demand, inventory, supplier costs). Dynamic pricing is potentially unlawful if: (1) it prices discriminates to exploit consumer vulnerability (e.g., targeting people with addiction or high debt as willing to pay more), (2) it misrepresents the product or pricing mechanism to consumers, or (3) it is part of collusive conduct (competitors setting prices through coordinated algorithms).
Real-world example: An Australian e-commerce platform adjusted prices based on a customer’s device type—customers on iPhones (indicating higher income) saw higher prices for the same products than Android users. The ACCC considered this algorithmic discrimination that potentially exploited consumers by misrepresenting uniform pricing, and the platform agreed to reform its algorithm and provide affected consumers compensation.
Fake Reviews and Algorithmic Amplification
The ACCC is investigating the use of AI to generate fake reviews, amplify positive reviews, and suppress negative reviews. This is misleading conduct under the Australian Consumer Law: consumers relying on reviews to make purchasing decisions are being deceived by algorithmically manipulated review signals.
The ACCC distinguishes between: (1) reviews posted by real humans (permissible), (2) AI-summarised or synthesised reviews (potentially permissible if accurately labelled), and (3) AI-generated fake reviews or algorithmic manipulation of review prominence (unlawful). A business that uses AI to rank reviews to amplify positive ones and demote critical ones, without transparent labelling, is engaging in misleading conduct.
Multiple Australian e-commerce platforms and hospitality businesses have been warned that algorithmic review manipulation violates consumer law. The ACCC has stated it is preparing enforcement action against repeat offenders.
Algorithmic Targeting and Vulnerability Exploitation
The ACCC scrutinises algorithmic targeting systems that identify vulnerable consumers and target them with high-margin products or manipulative claims. Unconscionable conduct under the Australian Consumer Law includes: targeting conduct that exploits a consumer’s apparent incapacity, age, health status, or financial stress.
If an algorithm identifies that a consumer has high debt, poor credit score, or financial vulnerability, and then targets them with expensive credit products with heavy marketing, the ACCC treats this as potentially unconscionable algorithmic conduct. The algorithm itself becomes evidence of intentional targeting of vulnerability.
A fintech platform was warned by the ACCC for using an algorithm to identify consumers in financial stress and aggressively market high-interest credit products to them. The algorithm’s targeting parameters (low income, high debt-to-income ratio, recent job change) demonstrated that the algorithm was designed to exploit vulnerability.
Algorithmic Recommendation Systems and Product Safety
The ACCC is investigating recommendation algorithms that promote unsafe products without appropriate warnings or context. For example, social media algorithms that amplify the visibility of cleaning products with high toxicity, or health claims that are scientifically unsupported or unsafe.
The legal issue: if an algorithm recommends a product without adequate product safety information, or amplifies unsafe products to consumers, the business deploying the algorithm shares liability for promoting the unsafe product. The algorithm is not a neutral delivery mechanism—it is an active choice to amplify certain products.
In 2024, the ACCC issued guidance clarifying that algorithmic recommendation of products comes with the same consumer law obligations as human sales: if a product is unsafe, or claims are misleading, using an algorithm to promote the product does not absolve the business of responsibility.
Australian Consumer Law Applied to AI-Enabled Conduct
Misleading or Deceptive Conduct (s.12CB, CCA 2010)
Conduct is misleading or deceptive if it is likely to mislead a substantial portion of the public about: the nature, quality, or safety of goods or services, the price or value, or consumer rights and guarantees. Algorithmic conduct falls squarely within this prohibition.
Examples: An algorithm misrepresenting product features to certain consumer segments, an algorithm hiding negative information from certain users, an algorithm creating artificial urgency by falsely suggesting limited availability, or an algorithm recommending products without disclosing conflicts of interest. All of these constitute potentially misleading conduct.
The ACCC’s position: the fact that a human did not manually make each misleading statement does not matter. If the algorithm is designed or deployed in a way that produces misleading outcomes, the business is liable.
Unconscionable Conduct (s.12CB, CCA 2010)
Unconscionable conduct includes conduct that is unfair, oppressive, or unreasonably difficult for a consumer to deal with, especially if it exploits vulnerability. Algorithmic targeting of vulnerable consumers, unconscionable pricing, or manipulative algorithmic design can all constitute unconscionable conduct.
Key factors the ACCC considers: whether the algorithm targets consumers based on vulnerability, whether the algorithm is transparent to consumers, whether it exploits information asymmetry, and whether the conduct is consistent with industry standards. An algorithm that does not meet these standards may be unconscionable.
False or Misleading Representations (s.12CB, CCA 2010)
False or misleading representations about goods and services are prohibited. This extends to representations made by algorithms. If an algorithm misrepresents product features, safety, or value to consumers, the business is liable.
Example: An algorithm recommending a product based on false representations of its effectiveness, safety, or environmental impact. Or an algorithm making recommendations without disclosing that the business receives a commission for those recommendations (failure to disclose a material conflict of interest).
ACCC Enforcement Actions and Warnings
The ACCC has escalated enforcement activity on algorithmic conduct. As of 2024, the ACCC:
Issued formal warnings to major retailers, travel platforms, and fintech companies about algorithmic pricing practices, fake review management, and algorithmic targeting.
Opened investigations into multiple businesses for alleged breaches of the Australian Consumer Law through algorithmic pricing, targeted advertising to vulnerable consumers, and fake review amplification.
Obtained agreements for algorithmic audits and reforms from several e-commerce platforms and financial services businesses that used algorithms in potentially unlawful ways.
Consulted on algorithmic governance standards, signalling that future enforcement will expect documented algorithmic governance, impact assessments, and audit procedures.
The ACCC has also made clear that it will pursue class action litigation against businesses using algorithms to systematically mislead or harm large numbers of consumers. Multiple cases are in investigation or pre-litigation phase.
What Businesses Must Do to Stay Compliant
Document Your Algorithms
Maintain clear documentation of what algorithmic systems you use, what data they use, what decisions they influence, and who approves deployment. Documentation becomes evidence in ACCC investigations: if you can demonstrate due diligence and governance, you reduce enforcement risk.
Conduct Algorithmic Impact Assessments
Before deploying an algorithm that affects consumer-facing decisions, assess: who does it affect, what decisions does it influence, what are potential harms or unfair impacts, and what safeguards are in place. Document this assessment. The ACCC now expects organisations deploying consumer-facing algorithms to have documented impact assessments.
Test for Discrimination and Fairness
If an algorithm makes decisions affecting consumer access, pricing, or recommendations, test whether the algorithm produces disparate outcomes across consumer groups. This testing creates a record of due diligence and can help identify algorithmic bias before the ACCC does.
Ensure Transparency
Consumers should understand, at a reasonable level, what algorithms are doing and how they affect them. This does not require explaining every mathematical parameter—it requires explaining in consumer-friendly language: what data is used, how the system works, and what the consumer can do to challenge outcomes.
Maintain Human Oversight
For high-impact algorithmic decisions (credit, access to services, pricing), have humans available to review and override algorithmic outcomes. The ability to request human review demonstrates consumer protection governance.
Audit and Governance
Conduct regular audits (at least annually for material algorithms) to check whether systems are producing fair, non-discriminatory, transparent outcomes. Document audit findings and remediation. This creates a compliance record that demonstrates reasonableness if issues arise.
Frequently Asked Questions
What is the ACCC’s position on algorithmic pricing?
The ACCC treats algorithmic pricing as subject to the Australian Consumer Law. Conduct is unlawful if: prices are set through collusion or price-fixing (even if automated), algorithms systematically charge different prices to exploit perceived vulnerabilities (unfair pricing), or pricing algorithms misrepresent product value or conditions. The ACCC has issued warnings and is investigating algorithmic pricing practices across e-commerce and travel sectors.
How does Australian Consumer Law apply to AI decision-making?
The Australian Consumer Law applies to conduct by AI systems used in business. Key prohibitions: misleading or deceptive conduct, unconscionable conduct, and false representations. If an AI system recommends unsafe products, misrepresents product features, or makes targeting decisions that exploit vulnerability, the business is liable. The consumer cannot sue the AI vendor; they sue the business deploying the AI.
What are the ACCC’s main enforcement actions on AI to date?
ACCC enforcement priorities (2023-2025): (1) Algorithmic pricing and demand-based pricing, especially in travel and e-commerce. (2) Fake reviews and AI-generated review amplification. (3) Targeted advertising that exploits vulnerability or misrepresents products. (4) Algorithmic recommendation systems that promote unsafe products. The ACCC is preparing enforcement action; multiple cases are in investigation phase.
Key Takeaway
Consumer protection law in Australia now explicitly applies to algorithmic conduct. The ACCC is actively investigating businesses using algorithms in ways that mislead, discriminate against, exploit, or harm consumers. Unlike copyright or privacy law (where guidance is evolving), consumer protection requirements for algorithms are clear and actively enforced.
Organisations deploying consumer-facing algorithms need to understand: algorithmic transparency and fairness are not optional governance considerations, they are legal compliance requirements. The ACCC expects documented governance, impact assessments, and audit procedures. Organisations that build these practices now reduce investigation risk, enforcement exposure, and reputational damage from algorithmic failures.
Navigating algorithmic compliance and consumer protection? Anitech helps organisations design algorithmic governance frameworks, conduct fairness and impact assessments, and establish practices that align with ACCC expectations and Australian Consumer Law. Contact us to establish algorithmic accountability that protects your business and your customers.