Artificial intelligence continues to dominate headlines across Australia and globally, with governments, universities and major corporations investing heavily in the technology. Yet despite rapid adoption and bold claims from tech leaders, many highly promoted AI breakthroughs have struggled to deliver meaningful real-world results.
According to Professor Daswin De Silva from La Trobe University’s AI Institute, the industry is increasingly defined not only by poor-quality AI content, but also by “nothingburgers” — developments that generated enormous hype but ultimately failed to meet expectations.
The Rise of AI ‘Slop’ and Hollow Automation
AI-generated content has become part of everyday life, from robotic emails to unhelpful customer service chatbots. In workplaces and universities, entire presentations and reports are now often produced with minimal human input, raising concerns about accuracy and quality.
Several high-profile failures have intensified scrutiny. Earlier this year, fabricated references generated by AI systems appeared in reports linked to consulting giant Deloitte and in papers submitted to a major AI research conference.
While the technology continues to improve rapidly, public confidence has become increasingly mixed. AI adoption is accelerating, but so are concerns about misinformation, reliability and overdependence on automated systems.
The term “AI slop” — describing low-quality machine-generated content — became so widespread that it was named Merriam-Webster’s word of the year for 2025. Even so, technology companies continue promoting AI as a productivity and cognitive enhancement tool.
Education’s AI Revolution Remains Elusive
AI Tutors Have Struggled to Deliver
Education was widely predicted to become one of AI’s biggest success stories. After concerns about AI-assisted plagiarism emerged in schools and universities, technology firms shifted focus towards personalised tutoring systems and automated learning support.
In 2023, OpenAI, Microsoft and the Gates Foundation backed online learning platform Khan Academy in developing Khanmigo, an AI-powered tutor designed to support students through guided learning.
Khan Academy founder Sal Khan described the platform as a tool that would “revolutionise education” during a widely viewed TED Talk.
Three years later, however, the anticipated transformation has not materialised. Khanmigo has since been discontinued, with Khan acknowledging that many students simply did not engage with it.
The system deliberately avoided giving students direct answers in an effort to encourage critical thinking and exploration. In practice, many users lost interest altogether.
Australian universities have similarly rushed to develop AI learning assistants, though there remains limited evidence showing measurable improvements in student performance or long-term educational outcomes.
Cognitive Offloading Raises Concerns
Another growing concern is “cognitive offloading” — the tendency for people to rely on AI tools to reduce mental effort.
Modern AI systems increasingly encourage extended interactions through techniques such as prompt chaining, where chatbots suggest follow-up tasks automatically. A student asking about a novel, for instance, may immediately be offered essay drafts, literary analysis or research summaries.
While this boosts engagement metrics and increases AI usage, critics argue it may also discourage independent thinking and weaken knowledge retention.
For educators already grappling with declining attention spans and digital dependency, the long-term consequences remain unclear.
AI Agents Are Falling Short in Workplaces
The Promise of Automated Management
AI agents — software systems capable of independently completing tasks across multiple platforms — were promoted as the next major workplace breakthrough.
Supporters argued they could reduce administrative workloads and streamline repetitive business operations.
Tech entrepreneur Jack Dorsey went further, suggesting AI agents could eventually replace middle managers by handling information flow and organisational coordination.
However, real-world experiments have exposed significant limitations.
AI startup Every trialled AI-generated line managers during workplace meetings, only to discover the systems produced endless discussion without reaching decisions or taking meaningful action. The exercise reportedly consumed millions of AI processing tokens while delivering little practical value.
Job Apocalypse Predictions Have Softened
During 2025, leading AI executives issued stark warnings about mass unemployment caused by automation.
Sam Altman and Dario Amodei both predicted severe disruption to entry-level employment markets.
Yet as their companies move closer to public listings, both leaders have moderated their rhetoric.
At the same time, major companies including Uber and Microsoft have reportedly found that maintaining large-scale AI systems can cost more than employing experienced human specialists to perform similar work.
Scientific Breakthrough Claims Also Under Scrutiny
AI’s role in scientific discovery has generated enormous excitement, particularly in fields such as materials science and medicine.
One of the most prominent examples was Google DeepMind’s GNoME project, which claimed AI had identified 2.2 million new material structures — an achievement the company described as representing “nearly 800 years’ worth of knowledge”.
But subsequent analysis by human researchers challenged those claims, arguing many of the proposed materials closely resembled existing ones or were the result of AI hallucinations presented as discoveries.
The episode added to growing concerns about verification standards in AI-assisted scientific research.
Separating AI Hype From Reality
Professor De Silva argues these highly publicised disappointments are likely to continue as competition intensifies within the AI industry.
While some AI developments are delivering genuine advances, the loudest claims often come from companies with strong commercial incentives to maintain public excitement and investor confidence.
For Australians navigating increasingly AI-driven workplaces, education systems and online environments, improving AI literacy may prove essential. Understanding the difference between technological capability and marketing hype will be critical to adopting these tools responsibly and effectively.
The challenge now is not whether AI will shape society, but whether governments, businesses and individuals can separate meaningful innovation from empty promises.
“Student. Subtly charming organizer. Certified music advocate. Writer. Lifelong troublemaker. Twitter lover.”
