Every generation has a definitive entry point into the workplace. For baby boomers, it was the fax machine. For millennials, it was the internet. For Gen Z, that is AI.
The youngest generation in Australia’s workforce can hardly dream of a working world without AI. Tools like ChatGPT, Gemini, and Claude are changing the way everyone works, and their use is being driven by Gen Z, digital natives with less experience with the business risks of AI.
That’s understandable. Like many previous technological breakthroughs, AI is being hailed as the ultimate tool for significantly increasing productivity and streamlining daily tasks. For lean teams, AI gives them the opportunity to punch above their weight and accomplish things that were only possible for large organizations just five or six years ago.
But while I’m an advocate of AI integration, there are also risks that small business leaders aren’t currently adequately mitigating. It’s the mismatch between AI’s trustworthiness and capabilities, especially among Gen Z workers.
Gen Z tends to exaggerate their AI capabilities
New research from RMIT Online and Deloitte Access Economies reveals hidden insights into each generation’s AI literacy and trust levels. Top line? Generation Z is the most common, technically Every generation is knowledgeable when it comes to AI, but they are the most likely to exaggerate their skills.
To the casual observer, Gen Z workers may seem like AI wizards. You can quickly generate prompts, write scripts, and spit out summaries. But just because you can type clever sentences into a chatbot doesn’t mean junior staff suddenly have the expertise to critically evaluate its output, spot hallucinations, question inherent algorithmic biases, or deal with complex copyright and privacy risks.
Here we get to the heart of the matter. difference between technical skills and judgement Skills using AI. Gen Z’s technical skills with AI are second to none (scored 3.7 out of 5 in our survey, compared to Gen X’s 3.2 and Boomers’ 2.9). However, technical skills are also twice as likely to improve compared to judgment skills.
To put all this into context, the truth is that almost half (47%) of Gen Z in Australia are actually still at entry level when it comes to AI. Yes, they are ahead of other generations (75% of baby boomers are entry-level), but a large portion of their working cohort still believes that their AI skills are better than they actually are.
Why AI literacy is a risk mitigation strategy
Overconfidence or misuse of AI has real costs for businesses in the form of rework, wrong decisions, and misplaced trust. Another study found that 95% of respondents admitted to experiencing an AI-related incident in the past two years, with an average financial loss of $800,000.
Large companies have legal, compliance, and risk teams to uncover these lapses. In an SMB, you are the risk team.
If your staff relies on AI for detailed research or software coding without having the underlying domain expertise to validate the output, you’re not necessarily automating productivity. If staff do not have the appropriate skills and As AI improves its decision-making skills, it only automates large-scale mistakes. True AI literacy is about more than input (prompts). It is also a critical evaluation of results.
Gen Z is at the top of the Dunning-Kruger curve
Essentially, Generation Z is at the peak of the initial confidence burst outlined by the Dunning-Kruger curve. Created by two psychologists in the late ’90s, the Dunning-Kruger Curve measures the level of confidence a person feels in a subject relative to the amount of time they spend learning about that subject. This illustrates the psychological phenomenon that those who know the least about a subject are often the most confident. When you first start learning about a new topic or technology, the learning curve is steep, and this creates initial confidence in the learner. But as they dig deeper, they finally begin to realize the complexity of what they are learning (and realize that they know less than they originally thought), and their confidence begins to decline. Only by studying more over time, your confidence will slowly increase again.
How to dismantle shadow AI in small businesses
So how can small business leaders solve this problem without hollowing out their team’s enthusiasm for technology? It has to be dismantled. Shadow AI – Unguided and unapproved trial and error practices by staff in carrying out daily tasks.
You don’t need a huge corporate budget or a strict compliance department to do this. SMB technology companies can implement a lightweight, judgment-focused literacy framework by taking three immediate steps.
First, shift your focus from tools to validation. Make it an absolute operating rule that no output generated by AI, including code, copy, or customer data analysis, leaves your business without documented human validation checks. Teach your junior team that their value is not in how fast they prompt the AI, but in how they criticize it.
Second, we move from trial and error to structured literacy. Don’t let your staff make decisions for themselves. Our report found that self-directed experimentation is the least effective way to build real AI capabilities. We offer short-term, structured and qualified upskilling that sets baseline standards for safe, ethical and effective use across your business.
Finally, bridge the stack between generations. Pair tech-savvy Gen Z employees with experienced, data-skeptical senior staff. Let junior teams drive speed of execution, but make sure your senior team applies professional guardrails, client context, and risk management.
The AI revolution is as much a human transition as it is a software transition. If you want to safely scale your small business, stop focusing on the technology stack and start investing in the human stuff.
Nic Cola is the CEO of RMIT Online.
Last updated: June 20, 2026 by Nic Cola
