Artificial Intelligence has become so much a part of our news cycles and everyday use, we don’t call it that anymore; it’s simplyAI. With Anthony Albanese announcing he’ll set up a new Office of AI, and the acceleration of data centres being built to facilitate new and more powerful AI platforms, questions are being asked about the risks in using AI to solve work tasks and how much we control or even understand the processes involved.
As more tech workers struggle with their future work prospects, and the chilling implications of unregulated AI reveal themselves, research is underway to identify the risks in using and trusting AI outputs. Lawyers, politicians, newspapers and airlines have all made spectacular blunders relying on AI hallucinations, and yet AI is seen as the unstoppable tech we must come to terms with.
What are the terms?
BehaviourWorks Australia is about to start a joint research project with Dhayani Kirubaharan from the Department of Accounting, funded through the Monash Business School AI in Research Seed Grant. It will be looking at how early career professionals develop their skills when they use AI for core work tasks before they have fully learned how to do those tasks themselves. AI has been around for a few years, and while research exists on students or experienced workers seeking productivity gains, we know much less about how AI affects people who are still developing their expertise.
Experienced workers exercise professional judgement through repeated practice, feedback, and sometimes making mistakes. They have learned over time how to collate and synthesise data and information to perform the various assessments and decisions required to do their work. For many in this category, AI can speed up different processes and increase productivity.
But early career workers are often under pressure to deliver quickly. Early career professionals in accounting, consulting, finance and analytics are using generative AI for core work tasks, often before they have fully developed the skills and judgement needed to do those tasks.
The main research question is whether using AI helps people build their judgement while they are learning, or whether it allows them to skip some of the harder learning.
Risk, trust and judgement
We are also looking at this through a risk culture lens. Knowing when to trust an AI response and when to check it is a form of risk assessment. The blunders listed above may have been avoided had proper checks been done on the AI outputs, but equally, those involved felt the pressure to produce and release their work publicly. Both using and restricting AI presents risks; hopefully not career-ending results.
The project has two aims:
● To document how early-career professionals use and scrutinise AI in core tasks
● To identify the individual and organisational conditions under which AI supports rather than displaces skill development, and
The project’s contribution is to reframe early career AI use as a question of skill formation and risk, with implications for management learning, organisational behaviour and policy.
We have seen what happens when key decisions that affect humans are not made by them (like the Robodebt scandal). We’re becoming aware of the risks of relying too heavily on technology that might fail and re-learning skills that can help if that happens.
As AI continues to expand into our daily lives, news feeds and workplaces, studies like this may show where our learning gaps are, and how to retain the skills we need to make sure AI creates good jobs rather than replacing them.
Related posts
