The latest round of hype about artificial intelligence could be genuine as the technology threatens - or promises - profound change.
People began to talk about an "AI spring" five years ago after a tech winter or downturn, and it's still blooming.
The most well-known tool is ChatGPT, which rapidly changed the rules for learning and assessment at school - and work - in a matter of months.
"Often AI has struggled with being over-promised, versus what it can actually do, particularly in the earlier versions," technologist Simon Elisha told AAP.
"The (AI) winters have really been about not so much the market investments in these technologies but the potential for them to be useful," he said.
Smaller rule-based models of the past half century have been programmed to drive search engines and email, do facial recognition, target marketing campaigns, or spot theft from bank accounts.
That's more mundane than the movie version of robots taking over the world, but humans are now confronted with persuasive digital material that can change the way people think.
Generative AI can create essays, computer code, music and images by drawing from vast swathes of information on the Internet.
"What's interesting about this generative AI approach is there's a fundamental difference - you don't have to train it to do what you need it to do," Mr Elisha said.
"Suddenly we've got this more general purpose approach ... it has way more emerging capabilities than we've ever seen."
As chief technologist for Australia, New Zealand and Oceania Amazon Web Services, Mr Elisha is also founder and host of the AWS Podcast, drawing from decades of experience in the industry.
"I've been doing this for over 30 years. I've seen things come and go," he said.
"I think this does have the potential to be really fundamentally changing the way a lot of things happen."
He said it was up to individuals, governments and organisations to use it responsibly.
"I think as humans we struggle when our thought processes are challenged and we have to change the way we do things."
Ethical AI usage is essential to protecting human rights and trust, according to the Australian Human Rights Commission.
The commission is concerned about privacy, the spread of false information, and bias or discrimination based on algorithms that could influence life-changing decisions in healthcare, banking and other services.
Modernising Australia's approach to AI will not be easy but it must be done, they have told the government.
A common struggle for parents and teachers is how to deal with generative AI churning out homework and solving complex maths questions. But Mr Elisha knows of one school that took a different approach.
"It told kids to do their homework with generative AI so that they understood what it can do, how it works," he said.
It's also popular as a "codebot", with software developers using rival firm GitHub's Copilot and other versions to create coding much more quickly.
But there is a degree of responsibility needed that must be built in from the start, to use code whisperers the right way, Mr Elisha said.
An American start-up called Anthropic has a company ethos of responsible usage.
"They talk about constitutional training - they've created a constitution about how the model should respond that sets frameworks about how to avoid bias and police yourself," Mr Elisha said.
"It's very early days, but the work that's been done around responsible AI and the ethics of AI is very applicable now."
