With AI transforming everything from healthcare to logistics to the funeral industry, the startup world is brimming with companies leveraging artificial intelligence.
For business leaders looking to innovate and grow, AI can appear as a wave of opportunity, but it can also appear as an insurmountable tsunami that you should embrace anyway because everyone else is doing it.
“But just because it's a buzzword doesn't mean your business needs AI,” warns Professor Ralph Zaburg of Adelaide Business School.
He advised business leaders to start by understanding AI and its potential applications.
“The first step is to become familiar with the technology and its applications,” he said.
“Most small and medium-sized businesses can find some value in analytics, but they need to consider whether AI can help.
“Fundamentally, it's up to the CEO to educate himself and decide if there's a place for him to do that.”
Mr Zurbrugg is Associate Dean for Research at the Adelaide Business School, University of Adelaide, and a leader in the field of business analytics.
His colleague, MBA director Gary Bowman, argues that despite the rapid advances being made in technology, “understanding AI” is essential for leadership.
“This is becoming a more important part of executive MBA curricula as business leaders need to understand the opportunities and threats that come with AI,” Bowman said.
“We provide a foundation in how AI works in practice, particularly in the context of data, while focusing on leadership requirements in a changing world to ensure our graduates are equipped to leverage AI responsibly.”
The Productivity Commission reported in February 2023 that “Australia needs to keep pace with technological developments to underpin future economic prosperity”.
The research finds that Australian businesses lag behind other OECD countries in their use of data analytics and AI, with only 6% of businesses using data analytics and even fewer using AI.
But before introducing AI into a business, whether through workflows or otherwise, executives and leaders need to consider the risks around security, ownership and appropriate use of data, Zurbrugg says.
“For example, every time an interaction occurs that uses Open AI technology via the API, that information is communicated to Open AI,” Zurbrugg said.
For many businesses, custom-built chatbots have provided an early introduction to AI integration because they are relatively easy to formulate and build, and data can remain in-house.
A breakthrough in the use of chatbots came with the global release of Open AI's Chat GPT just three months before the Productivity Commission report, providing a widespread practical introduction to the potential of AI.
Now it feels like everyone who uses a computer at work uses Chat GPT as an assistant, a research tool, or to frame a report.
Zarburg believes that as the use of AI becomes more widespread, ownership of data and ownership of ideas will be called into question.
“Certainly this is going to be a major legal issue in the future,” he said.
“When you look at AI-generated imagery, typically a co-ownership structure is being come up with, where the user can use it for their own purposes, but so can the creator of the AI.
“But when you do something more complex, like running a business and you’re feeding information into this big language model, you run into problems because the model isn’t your own. [their terms and conditions allow them to retain] Take that data and use it however you like.
“This highlights the important point that whatever communication you have, you need to make sure it's not a commercial secret or a secret to your clients or users.”
Customers owning their own data can also pose challenges for companies, not just in terms of consumer rights but also what data is collected, he said.
Woolworths is using its home-grown chatbot, Olive, to answer customer questions and resolve service issues, such as refunds for lost or damaged goods.
“If we're talking about bananas rotting, it doesn't matter where the data ends up,” Zarburg said.
“But you should think carefully before providing sensitive information through this mechanism.”
Meanwhile, Coles has deployed AI-managed cameras to track customers as they move through its supermarkets.
Though ostensibly to prevent theft, Zarburg said the system can also be used to collect data.
“The cameras are watching what you're looking at, what you choose to view or buy, and that's basic data as well,” Zarburg said.
“Did you implicitly consent to sharing your in-store browsing data?
“At this stage, those are kind of open questions that people aren't really thinking about, but they are questions to think about.”
Zarburg said that in the short term, AI will be deployed to replace repetitive automation tasks where the input and output of information can be “easily categorized” using narrower areas of artificial intelligence.
“When you go to a drive-through restaurant, instead of a human taking your order, you're going to move to a stage where your order is taken through an AI routine.
“But in the medium term, I think more complex strategic decision-making will be informed by AI.
“For example, when a company’s top management team is deciding whether to go ahead with a product, they weigh the various factors that need to be taken into account at that stage.
“While the decision-making process will still be driven by humans, it would not be surprising to see AI agents participating in top executive team discussions, offering insights and ideas on how to market its products or whether to invest in certain projects.
“I wouldn't be surprised if we see that type of use of AI in the next five years or so.”
Getting the most out of AI requires investment in data infrastructure, data cleaning, data integration processes, and data security. He emphasized the importance of data cleaning, saying biased data can lead to bias in algorithms, which can affect decision-making.
He pointed to Amazon and Google as high-profile examples of companies that managed to avoid long-term reputational damage.
In 2018, it was reported that Amazon had developed an AI algorithm to help it hire software developers and other technical staff, but the algorithm was trained on data from resumes the company received primarily from male applicants, reflecting the male-dominated tech industry.
Amazon's recruiting tools reportedly downgrade resumes that include the word “woman” or female college graduates.
Google also made headlines earlier this year when it released its AI tool, Gemini, which aimed to ensure that the images it generated reflected gender and racial representation, but reportedly showed unexpected biases, refusing to generate images of all-white groups and generating inaccurate historical images, such as a white, female US president.
“These are all due to biases in the data, not necessarily conscious biases, but biases that influence the decisions that they produce,” Zarburg said.
“You don't just train the algorithm, you have to make sure it works properly. AI is known to sometimes hallucinate and not give the answers you expect, so there's an element of risk management there.”
Looking again at the bigger picture, he said it is important to develop a governance framework to address the use of AI, especially given its rapid advancements.
This is a regulatory issue, but it also requires companies to act proactively, leaving leaders feeling unprepared. This is where an executive MBA can help bridge the gap between recognizing a looming problem and knowing what action to take, Bowman said.
“There is no substitute for face-to-face learning, especially in a programme like the Executive MBA, where so much value comes from exposure to senior leaders and world-class academics,” he said.
“With such an engaged and diverse population, we offer a learning experience that is hard to find anywhere in the world – and one that can never be replicated online.”
Find out more at Adelaide Business School's Executive MBA information session on 13 June.
