Your biggest AI risk may be your youngest employees - your greatest opportunity, your most experienced ones

The Beyond Prompting: Measuring the Generational AI Gap report, launched by RMIT Online with Deloitte Access Economics, surveyed over 2,000 Australian workers and found that 84% are already using at least one AI tool at their jobs, but only 7% have what the researchers classify as advanced AI literacy. More than half (54%) are still beginners.

The report introduces a new framework for measuring AI literacy across six domains: 

• Knowledge of AI systems.
• Practical skills. 
• Transferability across tools.
• Critical evaluation of outputs. 
• Ethical and legal awareness.
• Strategic judgement about when and how to use AI. 

As expected, younger workers are more digitally fluent. Gen Z and Millennials score an average of 3.7 out of 5 points, compared to 3.2 for Gen X and 2.9 for Boomers. More than three-quarters (76%) of Boomers sit at the beginner level, compared to 43% of Millennials, for example.

 
But the report also reveals that 21% of Gen Z workers overrate their AI abilities, compared to just 10% of Gen X and 8% of Boomers. Millennials sit at 17%. That overconfidence increases the risk of deploying AI without adequate judgement, overlooking ethical implications, or treating outputs as reliable without verification. 

The consequences can be significant. A study by the University of Melbourne and KPMG found that 56% of workers have made mistakes at work due to AI, and more than half have presented AI-generated content as their own. A survey of large companies in Australia and New Zealand by Infosys found an even more serious problem: 95% reported AI incidents over the past two years, with 18% losing between US$1 million and US$2 million and 8% losing more than US$2 million. 

 
On the other end of the scale, older employees with low AI literacy disproportionately hold top roles and shape how entire organisations approach the technology. 

As AI becomes more strategic, these leaders will not be able to delegate the interaction with the tools. They will need to use and understand their potential and limitations to make sound decisions.

Low AI literacy also negatively impacts salaries. The study estimates that closing the AI literacy gap among Boomers alone would generate a collective wage dividend of $3.1 billion. 

Upskilling half of all beginner-level workers to intermediate, across all generations, would create a dividend of $18.9 billion.

Another revealing finding across all generations is that workers are twice as likely to be more advanced in technical AI skills than in judgment skills.  

The biggest gaps are in critical evaluation (the ability to assess outputs for accuracy and bias) and transferability (the capacity to apply skills across different tools). Although these gaps exist for everyone, they are worse for younger workers.

Despite workers' perspectives, companies are not providing the necessary training. Only 48% of workers receive any formal AI training from their employer. The majority are left to figure things out on their own, with 31% relying on self-directed experimentation and 18% on trial-and-error learning.

The training oversights also follow a generational pattern. Over half of Boomers (52%) have not completed any AI training in the past year, compared to 22% of Millennials. Younger workers, however, are the most likely to say training is not worth the time (38% of Gen Z) or that their existing skills are sufficient. Again, the overconfidence is actively discouraging younger professionals from learning more.. 

The report suggests that the single most urgent step for companies is to improve AI literacy among leaders. The reason is that they set the strategies and create the conditions for responsible adoption. 

As for how to train the generations, the study proposes different approaches. 

For younger workers comfortable with AI, the priority is building judgment. That means developing critical thinking broadly, not just for AI use. Exposure to decision-making, to the consequences of errors, and to complex problems all build these skills. Short courses focused on critical evaluation and responsible use can help these employees. 

Many older workers already have the experience that produces better AI inputs and critical thinking. What they need is confidence and familiarity. The solutions are structured, hands-on training with practical use cases, combined with visible organisational support and access to tools.

All cohorts want training with concrete applications and clear rules for AI use.

If you need to train your teams, audit their current literacy gaps. Map where overconfidence and hesitancy are concentrated. 

Establish clear policies, including approved tools, guidelines for working with data, and measures to mitigate significant risks. 

Create targeted training for different generations, levels, or applications based on your audit results. Measure improvements over time by generation, role and starting level.

AI literacy is a judgment, a governance and a generational problem, all at once. The organisations that treat it that way will get the most from the technology and their people.