It\u2019s now a common experience to receive an AI-generated email that\u2019s robotic and hollow, or get a stream of useless chatbot responses when you just need some help from customer service.<\/p>\n
Worse yet, some people will dose up entire slide decks and project documentation with AI slop. Then there are the infamous cases of hallucinated references in a report by consulting firm Deloitte<\/a> and in dozens of papers at a top AI research conference<\/a> earlier this year. <\/p>\n The jagged frontier<\/a> of AI continues on its paradoxical route. On the one hand, there\u2019s increased adoption. On the other, increased concern about the limitations and risks posed by the technology.<\/p>\n While \u201cslop\u201d \u2013 in the context of poor quality AI content \u2013 was Merriam-Webster dictionary\u2019s word of the year for 2025<\/a>, tech executives are still keen for us to think differently<\/a>, to view AI tools as cognitive enhancers.<\/p>\n But the industry doesn\u2019t just facilitate slop. It\u2019s also rife with \u201cnothingburgers<\/a>\u201d. These are AI non-events that were wildly promoted and highly anticipated, but failed to deliver as expected in the real world.<\/p>\n An \u2018education revolution\u2019 that\u2019s still pending<\/p>\n Education, despite being one of the first victims of AI-driven plagiarism<\/a>, was touted to be on track for a technological revival through AI tutors and personalised learning.<\/p>\n In 2023, OpenAI, Microsoft and the Gates Foundation supported and funded the widely respected online learning non-profit Khan Academy to build Khanmigo, an AI tutor. Founder and CEO Sal Khan claimed Khanmigo would \u201crevolutionise education\u201d in a TED talk viewed by millions<\/a>. <\/p>\n Three years later in 2026 the revolution is still pending, while Khanmigo has been declared dead<\/a>. In Khan\u2019s own words, \u201cfor a lot of students, it was a non-event<\/a>\u201d. Khanmigo was designed not to provide direct answers in order to encourage learning and exploring. In practice, students didn\u2019t engage at all.<\/p>\n Many universities were quick to build<\/a> AI tutors, but hard data on actual contributions to student success remains to be seen.<\/p>\n Cognitive offloading<\/a> (using AI to reduce mental effort) and increased access to AI tools is a further contradiction in education. It\u2019s in the interest of AI providers to get users hooked, so they will write more prompts and use more tokens.<\/p>\n Prompt chaining<\/a> is a creative technique used by all most advanced AI models to predict potential next tasks as part of the response to the original prompt. A chatbot might answer a question about a classic novel, and then follow it with something like \u201cwould you like me to draft an essay addressing your question in the context of postmodern literature?\u201d.<\/p>\n While this increases AI model usage, it also increases cognitive offloading and begins to shape<\/a> how students think and acquire knowledge \u2013 which doesn\u2019t necessarily lead to better educational outcomes. <\/p>\n AI agents can\u2019t replace people<\/p>\n On paper, the arrival of AI agents<\/a> was a welcome addition to workplaces burdened with repetitive tasks involving multiple systems and data sources. More capable than chatbots, agents can operate on their own and make decisions to complete well-defined tasks.<\/p>\n It didn\u2019t take long for the rhetoric around AI agents to step up a notch when Twitter\u2019s billionaire co-founder Jack Dorsey (who had already laid off 40% of workforce at his fintech company Block<\/a>), claimed AI agents can replace line managers<\/a> and their function of routing information up and down the organisation hierarchy.<\/p>\n We don\u2019t have to look far for a non-event. When AI startup Every put this theory into practice, it quickly found out AI agent line managers in a team meeting resulted in endless chatter with no decisions or actions<\/a> that cost millions of tokens. <\/p>\n Leading AI CEOs Sam Altman<\/a> (OpenAI) and Dario Amodei<\/a> (Anthropic) were bold and loud in 2025 with their claims of \u201cwipeout of entry-level jobs\u201d and \u201cjob apocalypse\u201d.<\/p>\n But as their companies plan to go public in the next few weeks, both CEOs have walked back<\/a> these claims<\/a>. And reports from the CEOs of Uber<\/a> and Microsoft<\/a> show AI budgets are costing more than the salaries of human experts doing the same work. <\/p>\n AI isn\u2019t revolutionising science, either<\/p>\n Generative AI has also made its way into science<\/a>, much to the chagrin of those worried about the quality of what\u2019s being published in scientific journals. And there are non-events here, too.<\/p>\n The GNoME project was claimed an early win for Google DeepMind where AI was used to discover 2.2 million new material structures<\/a>. Google claimed this result was equivalent to \u201cnearly 800 years\u2019 worth of knowledge<\/a>\u201d. A few months later, the study was thoroughly examined by human experts<\/a> and dismissed as hallucinations that were poorly presented but also very similar to known materials.<\/p>\n These AI non-events will keep happening, and it\u2019s important for the public to keep paying attention when they do. While the capabilities of AI models are still improving at breakneck speed<\/a> and some are leading to genuine breakthroughs<\/a>, the loudest voices in the AI industry also have a vested interest<\/a> to keep up the hype.<\/p>\n For the rest of us, we must continuously brush up on AI literacy, to build awareness of what separates hype from reality. Only then can we adopt AI responsibly, with a clear view of both the risks and the benefits.<\/p>\n","protected":false},"excerpt":{"rendered":"It\u2019s now a common experience to receive an AI-generated email that\u2019s robotic and hollow, or get a stream…\n","protected":false},"author":2,"featured_media":701020,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[256,254,255,64,63,105],"class_list":["post-701019","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-au","tag-australia","tag-technology"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/posts\/701019","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/comments?post=701019"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/posts\/701019\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/media\/701020"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/media?parent=701019"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/categories?post=701019"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/tags?post=701019"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}