Show Notes
There have been several noteworthy developments in the world of AI this week, and in today's episode, we share our take on the good, the bad, and the WTF. We start off by discussing exciting developments at The Human Technology Institute and how their new project, The Future of AI Regulation in Australia, aims to help laws and regulations catch up with language learning models like ChatGPT. Next, we reflect on the tragic news of the first death directly linked to a chatbot, how little the public truly understands about these technologies, and what can be done to minimize these types of risks in future. Our conversation then delves into the proposed moratorium to halt the development of more sophisticated versions of ChatGPT and key concerns that are being raised around this topic. We also spend some time unpacking the problematic source behind this petition, namely, The Future of Life Institute, before discussing what experts like Emily M. Bender And Gary Marcus have to say about the topic. To learn more about this week in AI, including our top takeaways, be sure to tune in today!
Key Points From This Episode:
The Human Technology Institute's new project: The Future of AI Regulation in Australia.
Reflections on the first tragic death directly linked to a chatbot.
Our thoughts on the petition for a moratorium to halt the development of more sophisticated versions of ChatGPT.
The problematic elements of the proposed moratorium, including its lack of boundaries and guardrails.
A reminder of the controversial origins of the Future of Life Institute and why the fact that they are funding this petition should act as a warning sign.
An overview of the discourse around this topic and our key takeaways.
Why laws and regulations have failed to keep pace with the language learning models like ChatGPT.
How the Human Technology Institute's new project aims to improve regulation, ensure more equitable distribution of wealth, and minimize the potential harm of these technologies.
Links Mentioned in Today's Episode:
'Man ends his life after an AI chatbot 'encouraged' him to sacrifice himself to stop climate change' 'The first known chatbot associated death' 'Elon Musk Signs Open Letter Urging AI Labs to Pump the Brakes' 'The Future of AI Regulation in Australia' Ed Santow on LinkedIn University of Technology Sydney (UTS) Human Technology Institute (HTI) Emily M. Bender Emily M. Bender on LinkedIn Emily M. Bender on Medium 'On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?' 'Policy makers: Please don't fall for the distractions of #AIhype' The Future of Life Institute Natalie Rouse on LinkedIn Dr Kobi Leins Dr Kobi Leins on LinkedIn Dr Kobi Leins on Twitter Eliiza
More from AI Australia Podcast
Most organisations treat AI governance as the thing that slows innovation down. Luli Adeyemo believes the opposite is true.
As part of our Aussie Start-Up Series we talk to Jo Thomas, founder of Enrola an AI sales agent platform helping B2C sales teams convert more leads into customers. Jo and her co-founder have built an AI sales agent that her own customers now refer to by name, and love. In this episode, she breaks down the maths behind why a $80,000-a-year employee might only be delivering $15,000 of real value, why AI works best as an amplifier for humans rather than a replacement, and where she draws a hard line: no AI should be helping someone choose their health insurance or their degree. A clear-eyed look at where automation earns its place, and where it doesn't.
Everyone's convinced AI is coming for engineering jobs. Tom Goerke, National President of Engineers Australia, thinks that fear is almost exactly backwards, and he's got 19th-century coal miners to prove it. In this episode, Tom unpacks Jevons' paradox: the counterintuitive economic pattern where making something more efficient doesn't reduce demand for it, it explodes it. He walks through what that means for software engineers right now (spoiler: more code is being written than ever, and more bugs with it), why "coders are the typists of the '70s," and why the real crisis isn't job losses, it's a widening skills gap between engineers who've mastered AI and everyone else. We also cover the one stat that stopped us in our tracks: 66% of engineers have had zero formal AI training from their employer.
Lee Hickin has spent 30 years at the frontier of technology, from IBM and AWS to serving as CTO of Microsoft Australia. Now, as Executive Director of the National AI Centre, he's asking a confronting question of every boardroom in Australia: when the risk and the opportunity depends on AI, who in the room is actually the expert?
Before ChatGPT changed everything, Arvind Jain was already four years deep into solving a problem every knowledge worker knows but no one had fixed: why is it so impossible to find anything at work? In this episode, the CEO of Glean, the enterprise AI platform used by some of the world's most iconic companies, takes us inside the founding story that started not in a garage, but in the wreckage of hypergrowth. When his previous company grew from 500 to 2,000 people in a single year, things started to break. Software releases stalled, teams lost each other, and the lowest-scoring question in their employee survey? Do I have the information I need to do my job? That moment of failure became a billion-dollar idea. Arvind opens up about his time in the early days of Google, co-founding Rubrik, and then walking away from success, again to build something that didn't exist yet. This one is for anyone who's ever wasted an afternoon searching for a file that should have taken 30 seconds to find.