The Governance Institute of Australia issued the warning on Wednesday as it released a policy paper to guide companies in the use of artificial intelligence.
The white paper comes almost one week after the federal government released voluntary guidelines on AI use, and a public consultation into mandatory restrictions that should apply in high-risk settings.
Governance Institute of Australia chief executive Megan Motto launched the document, saying it had been developed after three roundtables with AI specialists, including representatives from the CSIRO, QUT and National Artificial Intelligence Centre.
Using the advanced technology, she said, had become essential for businesses to remain competitive but rules about its use could be vague and confusing.
"Good governance is essential to ensure that AI technologies are developed and used in ways that are transparent, fair and aligned with societal values," she said.
The white paper expands on the 10 voluntary AI guidelines launched by Industry Minister Ed Husic on Thursday, which included establishing an AI strategy and training, identifying risks and monitoring AI tools.
The institute's policy details considerations for each step, including ethical considerations, the types of bias possible in AI systems, and privacy warnings.
National Artificial Intelligence Centre director Stela Solar, who collaborated on the guide, said research consistently showed Australian businesses were slow to adopt AI technology and were cautious about its risks.
But she said that did not mean business leaders were not interested in safety unlocking its potential.
"Australia is the number one most nervous country in the world about AI but, in what I hope will be a great surprise to you, is that we are also the most curious," she said.
"I like to call it world-leading AI curiosity."
Clayton Utz partner Simon Newcomb said the government's voluntary guidelines and proposals for mandatory rules proved there were serious risks in using AI and existing Australian laws were not equipped to deal with them.
Problems had emerged with the misuse of copyright material to train AI, he said, its use to create deepfakes and phishing campaigns, and with bias in fields such as recruitment.
But Mr Newcomb said businesses needed to persist and establish clear strategies for its safe use.
"There's enormous opportunity with AI and also some significant risks," he said.
"The trick is here not to shut it down – I wouldn't be advocating that to anyone."
Research from the Tech Council found that generative AI could contribute up to $115 billion per year to the Australian economy and create up to 200,000 jobs by 2030.
