Logo
FrontierNews.ai

10.5 Million Australians Use ChatGPT, Yet OpenAI's Trust Is Collapsing. Here's Why.

OpenAI is experiencing a stark disconnect in Australia: millions of people use ChatGPT regularly, yet trust in the company is rapidly eroding. According to Roy Morgan's latest Trust and Distrust rankings, OpenAI has fallen three places to become Australia's 16th most distrusted brand, even as approximately 10.5 million Australians aged 14 and older actively use ChatGPT. This paradox reveals a growing tension between the utility of AI tools and public anxiety about the companies behind them.

Why Is Trust Declining When Usage Is Soaring?

The disconnect between adoption and trust reflects deeper concerns about AI's societal impact. Roy Morgan CEO Michele Levine noted that distrust in OpenAI has grown quickly, but not because of a single scandal or major incident. Instead, Australians are increasingly worried about what AI companies represent and what they might do next. These concerns span multiple dimensions, creating a cloud of uncertainty around the industry.

The specific worries driving distrust include:

  • Privacy Concerns: Users fear how their data is collected, stored, and used by AI companies to train future models.
  • Ethical Questions: Australians worry about the ethical implications of AI systems making decisions that affect people's lives.
  • Profit Motive: Distrust stems from concerns that companies prioritize financial gain over user welfare and societal benefit.
  • Environmental Impact: The enormous data centers required to power AI systems consume vast amounts of energy, raising sustainability questions.
  • Loss of Cognitive Autonomy: Some Australians fear that relying on AI might diminish our ability to think independently.

Notably, these concerns are often anticipatory rather than reactive. Many Australians are forming opinions about AI companies before experiencing problems firsthand, suggesting that the narrative around AI risk is shaping public perception as much as actual company behavior.

How Does ChatGPT Compare to Other AI Platforms in Australia?

ChatGPT dominates the Australian AI landscape by a significant margin. Approximately 45 percent of Australians aged 14 and older use the platform, more than double the adoption rate of Google Gemini at roughly 5 million users and well ahead of Microsoft Copilot's estimated 4 million users. Smaller players like Canva Magic Studio reach about 1.4 million Australians, while Anthropic's Claude reaches approximately 777,000 users.

Overall, around 13.6 million Australians aged 14 and older now use at least one AI platform in an average four-week period, demonstrating how rapidly AI adoption has accelerated across the country. This widespread usage spans age groups more evenly than many might expect. AI adoption is highest among Australians aged 25 to 34 and 35 to 49, with 74 percent and 72 percent using at least one AI tool respectively. However, younger demographics are not leading the charge as some might assume. Usage among Australians aged 18 to 24 sits at 68 percent, while 66 percent of those aged 14 to 17 use AI tools.

Can Distrust Coexist With Continued Usage?

History suggests the answer is yes. Distrusting a company does not necessarily mean people stop using it. Optus and Telstra rank first and sixth on Roy Morgan's most distrusted list, yet both remain woven into everyday Australian life through phone bills and services. Similarly, Coles and Woolworths occupy fourth and fifth place on the distrust rankings, yet Australians continue shopping at these supermarkets because convenience, habit, and lack of alternatives often outweigh trust concerns.

The same dynamic is now playing out with AI. Millions of Australians find enough value in ChatGPT to keep using it, even as concerns about OpenAI continue to grow. Using a brand is not the same as trusting it. Sometimes the product is useful enough that people are willing to set their reservations aside. The difference with AI, however, is the speed at which this uncomfortable relationship has developed. Telcos, supermarkets, and banks had decades to become fixtures in Australian life. ChatGPT has managed it in only a handful of years.

Roy Morgan's data shows that 65 percent of Australians distrust at least one company, compared with only 47 percent who trust one. Since mid-2020, Australians have actually been more distrustful than trusting overall, suggesting that skepticism toward corporations is a broader cultural trend rather than an AI-specific phenomenon.

What Does This Mean for OpenAI and the Broader AI Industry?

The Australian trust data offers a cautionary signal for OpenAI and other AI companies. Building massive user bases is one challenge; maintaining public confidence is another. The company's rapid rise in distrust rankings suggests that as AI becomes more capable and more integrated into daily life, public scrutiny will intensify. Companies in the space cannot rely on utility alone to sustain their social license to operate.

The broader AI sector is also consolidating around trust and distribution challenges. Gamma, a presentation software company that has crossed 100 million users, recently acquired Lica, an AI design startup, to build a research lab focused on how AI models communicate across formats and audiences. This move reflects a strategic recognition that distribution and research depth are both necessary to compete as larger players like OpenAI expand their product offerings. When OpenAI acquired NextSlide earlier in August 2026, it signaled that frontier AI labs are moving native AI capabilities directly into their flagship products, forcing independent players to either specialize, sell, or invest heavily in differentiation.

For OpenAI specifically, the Australian trust data underscores a challenge that will likely extend beyond that market. As AI systems become more powerful and more widely deployed, questions about privacy, ethics, environmental impact, and corporate accountability will only grow louder. The company's ability to address these concerns transparently and substantively may ultimately matter more than the raw capability of its models.