Australian regulator warns of personalized pricing and shopping bots that could collude against consumers

By The Nightly (AAP) · October 17, 2026.

The Australian Competition and Consumer Commission (ACCC, the country's consumer watchdog) warned a parliamentary committee on AI on Friday that the technology can be used to charge some customers more than others. It also cautioned that shopping bots could steer users away from the best deals.

According to David Abkiewicz, the ACCC's head of digital markets, the real harm arises when personalized pricing becomes “extremely granular.” In that case, he said, it could pin down each person's willingness to pay entirely, or rely on data revealing that the customer is vulnerable.

Abkiewicz himself acknowledged that offering different prices to certain groups does not breach Australian consumer law. He added that other jurisdictions, including the U.S., are starting to restrict surveillance-based pricing. According to the ACCC, there is currently no straightforward way to act against companies that exploit vulnerable customers with tailored prices.

The second concern is “agentic commerce”: consumers handing off purchases to an AI agent, that is, a program that acts on their behalf. According to the article, consumer advocates fear these agents could have built-in brand preferences and might even collude with one another to steer shoppers away from the best deals. Abkiewicz said the regulator will continue to monitor whether these agents act in the consumer's interest.

Jayde Richmond, director of the ACCC's National Anti-Scam Centre, noted that AI allows scammers to operate at scale and to automate and personalize their outreach. She cited voice cloning, deepfake videos and possible use to manipulate search results, though she said the latter is not yet being reported in large numbers.

Abkiewicz further warned that, without intervention, a digital economy increasingly shaped by AI tends to produce dominant players that are hard to dislodge. He added that we are still at a very early stage. So far, the ACCC's most visible case is its lawsuit against Microsoft, which it accuses of misleading 2.7 million Australians about pricing when it integrated Copilot into a consumer product. That is the regulator's allegation, and the article does not treat it as proven.

Our take: the key finding is not that personalized pricing exists, but that the law has no clear answer to it. A price tailored to what each person is willing to pay is not a technical flaw in AI. It is the effective execution of a commercial goal, which is why the right path is regulation, not waiting for the market to self-correct.

The shopping-agent issue is subtler. The promise is an assistant that works for you and scours the market better than you can. But, as with any intermediary, the decisive question is who funds it and who takes a commission. If the agent is paid for by the seller or a platform, it inherits their conflict of interest. The suspicion of collusion between bots is best treated as a hypothesis regulators are watching, not as a fact documented in this article.

Overall, these practices are the short-term dark side of a technology with great potential: the same capability that allows prices to be fine-tuned against the consumer can, if well governed, be put on their side and make finding the best option cheaper. As with other transitions, the challenge is one of rules and transparency. Requiring shopping agents to declare their neutrality and make their revenue model visible would be a reasonable first step, and an evergreen one: it will still be the right question once these assistants are part of everyday life.