Lately there has been much talk, especially among politicians, about productivity. And rightly: the growth of labor productivity in Australia is at its lowest level in 60 years.
To address this problem, Australia Prime Minister Anthony Albanese has called a round table about productivity for next month. This will coincide with the publication of a provisional report of the Productivity Commission, which analyzes five reform pillars.
One of them is the role of digital data and technologies, including artificial intelligence (AI).
This will be an encouraging news for the technological and business sectors, which have enthusiastically promoted the benefits of AI for productivity. In fact, the Australian Business Council said last month that AI represents the greatest opportunity in a generation to improve productivity.
But what do we really know about how AI impacts productivity?
What is productivity?
In simple terms, productivity is the amount of products (goods and services) that can be generated from a certain amount of inputs (such as labor and raw materials). It is important because greater productivity usually translates into a better standard of living.
In fact, productivity growth has represented 80% of income growth in Australia during the last three decades.
Productivity can be analyzed at the individual, organizational or national level. Your individual productivity refers to how efficient you manage your time and resources to complete tasks. For example, how many emails do you answer in an hour? How many products do you review in search of defects per day?
Organizational productivity refers to how effective an organization is in the fulfillment of its objectives. For example, in a research organization, how many high quality scientific articles are produced?
See: The therapist of AI will attend you now: Can chatbots really improve mental health?
National productivity is the general economic efficiency of a country, usually measured as GDP per hour worked. It is, in effect, an aggregation of the other two. However, it is notoriously difficult to track how changes in individual or organizational productivity are reflected in national GDP per hour worked.
AI AND INDIVIDUAL PRODUCTIVITY
Initial investigations that explore the relationship between AI and individual productivity yield mixed results.
A 2025 study on AI and productivity included 776 product professionals from the multinational Procter & Gamble. The study showed that those who used the two -people team without it.
A similar study in 2023 with 750 Boston Consulting Group consultants revealed that the tasks completed 18% faster with generative.
Another 2023 article documented the use of a generative AI system by 5,200 customer service agents in a software company from the Fortune 500 list. A 14% increase in the problems solved per hour was observed, and in the case of agents with less experience, productivity increased by 35%.
However, AI does not always improve individual productivity.
A survey of 2,500 professionals revealed that 77% felt an increase in their workload by using the generative AI, and 47% said they did not know how to take advantage of it. The main barriers indicated were the need to verify or correct results, the lack of skills in AI and unrealistic expectations about their capacity.
A recent study of the CSIRO examined the daily use of Microsoft 365 Copilot among 300 employees of a government entity. Although most reported improvements, a considerable proportion (30%) did not. Even those who did improve their productivity expressed expected major benefits.
AI AND ORGANIZATIONAL PRODUCTIVITY
It is difficult – if not impossible – attributing changes in organizational productivity only to the implementation of AI, since multiple social and structural factors intervene.
Even so, the Organization for Economic Cooperation and Development (OECD) estimates that the benefits of traditional AI (such as automatic learning applied to specific tasks) range between 0% and 11% at the organizational level.
A 2024 report cites independent studies that record increases in productivity thanks to AI in Germany, Italy and Taiwan.
On the contrary, an analysis of 2022 over 300 thousand companies in the US did not find a significant correlation between the adoption of AI and productivity, although it did find it in technologies such as robotics and cloud computing. This could be due to the fact that AI has not yet had a broad impact, which is difficult to isolate its effect, since it is never implemented by itself.
In addition, the increase in productivity provided by AI can be compensated by the additional labor required to train or operate these systems. An example is Amazon’s “Just Walk Out” technology, publicly launched in 2018 to automate shopping shopping.
According to reports, the system depended on about a thousand workers in India to perform quality control tasks. Amazon has described these reports as “erroneous.”
In more general terms, let’s think of the millions of people who are probably involved in the task of labeling data to train AI models.
AI and productivity in Australia
In the Australian country the panorama is even more uncertain.
Until now, AI has not had a visible impact on national productivity. Some argue that technological advances require time to translate into productivity improvements, since companies need to adapt, implement infrastructure and train their staff.
However, this is not guaranteed. For example, although there is consensus in which the Internet improved productivity, the effects of mobile phones and social networks are more debated. Its benefits are more noticeable in some sectors, such as entertainment, than in others.
Find out: Study alert to the difficulty of erase sensitive data in artificial intelligence
Productivity is not just doing things faster
The common narrative suggests that AI improves productivity by automating routine tasks, releasing time for creative activities. But this vision is simplistic.
That you respond your emails faster does not mean that you over time to go to the beach. The more emails you send, the more you will receive. The cycle does not stop.
Faster is not always better. Sometimes we need to decelerate to be more productive. It is in that space where the great ideas are born.
Imagine a world in which AI not only expedite tasks, but also help us to stop deliberately to think, create and innovate. That is the true opportunity not yet exploited from artificial intelligence.
*Jon Whittle is director of Data61, Csiro.
This article was originally published in The Conversation
Do you like photos and news? Follow us on our Instagram
