Australia is at risk of falling behind in the rapidly advancing field of artificial intelligence (AI), a concern increasingly echoed by industry leaders and academics. Professor Toby Walsh from the University of New South Wales describes the current situation as the “largest gold rush in the history of capitalism,” warning that Australia is missing out. This issue extends beyond financial gains; it threatens the country’s overall prosperity and its ability to transition to an AI-enabled economy.
The urgency of this message was underscored during a recent Senate Select Committee on Adopting Artificial Intelligence, where experts highlighted Australia’s sluggish progress compared to its global peers. Professor Anton Van Den Hengel, Chief Scientist at the Australian Institute of Machine Learning, emphasized that Australia has been notably slow in making the necessary transition. While some may view the call for AI investment from researchers as self-serving, the stark contrast in funding between Australia and other nations supports their concerns.
Professor Walsh pointed out that since 2017, Australia has invested less than a tenth of Canada’s budget in AI research and development. This discrepancy is even more striking when considering that countries with comparable or smaller gross domestic products (GDP) have allocated billions to AI initiatives. One proposed solution from Professor Van Den Hengel involves developing a homegrown Large Language Model similar to OpenAI’s ChatGPT, rather than relying on imported technology. He argues that active participation in AI development is crucial for Australia to influence global advancements in this critical area.
Cookie Consent and Privacy: A Disparity in Digital Experiences
Australians spend significantly more time managing their privacy settings on websites compared to their European counterparts, a frustration familiar to many internet users. According to research from the Consumer Policy Research Centre, Australians average roughly two minutes per site, whereas Germans achieve the same results in just 3.1 seconds. This disparity stems from the European Union’s stronger consumer privacy laws.
Australia has been working on overhauling its Privacy Act for several years, with changes in progress since a review initiated by the Morrison government in 2019. In May, Attorney-General Mark Dreyfus announced that the bill’s introduction would be accelerated to August, in part as a response to the government’s policy on violence against women. Despite these promising signs, there remains a cautious optimism among advocates, akin to the anticipation for a long-awaited album release.
Consumer advocacy group CHOICE recently delivered a petition with nearly 33,000 signatures to Prime Minister Anthony Albanese, Treasurer Jim Chalmers, and Dreyfus. The petition urges the government to implement reforms that would compel businesses to act fairly and responsibly when handling consumer data. Kate Bower, CHOICE’s consumer data advocate, emphasized that the issue extends beyond the annoyance of cookie management. She highlighted a culture of rampant data over-collection in Australia, leading to significant breaches and unfair practices such as price discrimination and manipulative marketing.
Despite the pressure to act swiftly, Dreyfus was vague about the timeline during a National Press Club address, expressing hope that the bill would be presented to parliament later in the year.
Chatbot Misinformation: A Growing Concern
Misinformation monitors at NewsGuard have found that leading generative AI platforms, including OpenAI’s ChatGPT-4, xAI’s Grok, and Google’s Gemini, frequently fail to provide accurate information about recent events. Specifically, they were tested on their responses to the recent shooting incident involving Donald Trump. Alarmingly, 57 percent of the time, these platforms did not debunk false claims, with only one consistently providing accurate information.
The proliferation of misinformation, particularly on social media, exacerbates this issue. Following the rally where the shooting occurred, social media platforms saw a surge in false narratives, including claims that the attack was ordered by Joe Biden or staged by Trump himself. On X (formerly Twitter), there were 308,000 mentions of “staged” and 83,000 mentions of “inside job” in the immediate aftermath.
Interestingly, the AI models more often provided non-answers rather than outright falsehoods. Of the inaccurate responses, only a fifth repeated false information, while the majority deflected the questions, claiming an inability to discuss politics or similar topics.
As AI becomes increasingly integrated into consumer technology and search engines, the challenge of ensuring these systems provide reliable information grows more critical. The current shortcomings underscore the need for continued vigilance and improvement in AI accuracy and reliability.
