�Ӱ�ֱ��

On Friday, 11 April, government and industry representatives converged at the AIML building in Adelaide for a lively and though-provoking discussion on AI’s continuing role in the future of business.

The event, titled Generative & Agentic AI for Business, was hosted by the (NAIC), and the (Ai Group), a national employer organisation representing traditional, innovative, and emerging industry sectors.

Speakers at the event included AIML Chief Scientist, Professor ; AIML Professor and interim director of the Responsible AI Research (RAIR) Centre ; AIML adjunct member and Managing Director of Dr Jamie Sherrah; Jannat Maqbool, Manager for Industry Programs at the (ARM) Hub AI Adopt Centre; and Marshall Cowan, CEO of .

Rita Arrigo, NAIC’s Strategic Engagement Manager, was on hand as MC of the event with support from Estha van der Linden of Ai Group.

Michael Brown, South �Ӱ�ֱ��’s first Assistant Minister for Artificial Intelligence and the Digital Economy, kicked things off with a welcome speech where he noted the work being done by the South �Ӱ�ֱ��n government to promote continuing research and development in AI.

“I’m very pleased as the Assistant Minister to have the opportunity to advance the State Government’s focus on AI strategy for our state,” he said in his speech.  “Aligning with our economic objectives, a South �Ӱ�ֱ��n AI Strategy is now being developed to grow investment in AI, to strengthen AI capabilities, to boost AI skills, to raise productivity and efficiency, and to develop sovereign AI capability and resilience.”

“By leveraging AI to analyse and utilise data, we can create more successful businesses, and attractive conditions for more companies to set up in our state.”

Marshall Cowan of Clevertar addressed the FOMO, or fear of missing out, factor in businesses’ adoption of AI.

“Business is a game, a competitive game,” he said. “Everyone from boardrooms to investors now expect an AI strategy. And no one wants to be left behind.”

“But fear-based adoption often leads to poor outcomes. AI can get you 80% of the way there, but the final 20% is where risk and real engineering live.” Cowan recommends that businesses interested in developing an AI strategy develop a clear, knowledge-based plan that incorporates input from the entire organisation. He also suggests developing a “top down” strategy focusing on AI literacy and well-defined, measurable value across the organisation. 

Addressing the concerns of business and the public

In a talk titled, “It's Stochastic Parrots all the way down,” Professor van den Hengel discussed recent advancements in AI technology and the public’s reaction to them. 

“Stochastic parrots’ is a reference to the idea that chatbots are just parroting our knowledge back to us,” said Professor van den Hengel. “Until 3 years ago, chatbots were rubbish. Anything that responded usefully was human. [But these improvements] have caused people to overestimate the short-term implications of AI [such as job losses] while underestimating the long-term impact.”

“The impending immediate doom we often hear about is overblown.”

Professor van den Hengel also called for greater investment in �Ӱ�ֱ��n-based AI and AI initiatives.

“�Ӱ�ֱ�� should provide opportunities here so that for every 1 person that leaves, we get 1 who arrives,” he said. “AI companies around the world are full of Aussies. As a nation, we have to think about how are we providing opportunities to retain and retract the people who’ll drive our transition to a more advanced economy.”

In his talk, Dr. Jamie Sherrah discussed the risks many �Ӱ�ֱ��n businesses associate with adopting AI, including AI hallucinations, legality, liability, and price. Dr Sherrah noted that one way to improve �Ӱ�ֱ��’s standing in the global AI race – and to address some of the concerns around legality and liability - is to create more �Ӱ�ֱ��n AI companies that keep �Ӱ�ֱ��n data in �Ӱ�ֱ��.

“Most AI tools are hosted in the U.S., which means your data is being stored outside of �Ӱ�ֱ��,” said Dr Sherrah. “Many �Ӱ�ֱ��n organisations are obliged to conform to the �Ӱ�ֱ��n Privacy Act of 1988, which requires that data disclosed to a third party (e.g. an AI tool) overseas be guaranteed the same legal privacy protections as in �Ӱ�ֱ��.”

Dr. Sherrah’s solution was to develop Hippo, a chatbot application similar to ChatGPT or Perplexity, but one based in �Ӱ�ֱ��. According to Dr Sherrah, Hippo is committed to data privacy and ensures that customer data is never used for AI model training.

“In addition, Hippo operate 10 times faster than ChatGPT – but that’s probably because we don’t have as many users yet,” Dr Sherrah said to laughter from the audience.

For more information on Dr Sherrah’s work with Hippo, please read his article, “” on AIML’s Medium page.

AI and the future

AIML Professor Javen Qinfeng Shi took to the podium to discuss the role of intelligence and causality in AI.

“Would you consider a book intelligent? How about a calculator?” he said during his talk. “The ability to solve new problems [is] kind of a sign of intelligence.” 

“[But true intelligence is the] essence of what it is to learn, which I call wisdom function.”

With causality, Professor Shi observed that there are two main research questions. One is what he calls the ‘forward question’ or, if you change something what's changing with it? The second question is if you have a dream for a future that you really want, what are the interventions not within your power that could prevent you from achieving it? The list of constraints could include talent, resources, money, or the amount of time one has to shape the future.

“What's really important is the future,” he said.

Professor Shi went on to note that there were two types of AI: one to help businesses and industries grow and expand, and a second to help humanity become better.

“Now is the time to think about what we need to do to become better, more compassionate humans. If you want to…empower you in the future and your kids, [your] grandchildren in the future, [please join] and support us” through AIML’s work with RAIR.

For more information on Professor’s Shi’s work on causality and AI, please read his article, “” on AIML’s Medium page.

Expanding on Professor Shi’s point about the first type of AI, Jannat Maqbool discussed effective ways to use AI to bridge the gap to commercial delivery. 

“Developing technology assessments and road maps is the first step,” she said. She noted that these assessments must address important issues such as governance and security and be understood by non-technical users.

Maqbool discussed the current industry shift of moving from ‘software-as-a-service’ to ‘service-as-a-software,’ where AI systems generate outcomes such as RFPs, contracts, and other outputs. She also stressed the importance of “building a zero-waste knowledge organisation.”  

“Let’s not store data,” she said.  “Let’s use it to generate outcomes and use AI agents to boost productivity.”

The event concluded with a lively panel discussion where participants discussed best practice strategies to implement AI into business operations across the country, and shared stories of their own efforts to introduce AI-enabled tools to various corporations.