Deirdre Caren

Federal Roundtable Risks Missing the $116 Billion AI Prize The Australian Treasurer’s Economic Reform Roundtable begins today, and it’s already overlooking the single biggest lever for productivity this century: Artificial Intelligence. From the steam engine to the internet, every major leap in productivity has been driven by technology. AI is the next, and perhaps the most transformative, step. The Productivity Commission estimates Australia could gain $116 billion if we accelerate AI adoption. Yet when Treasurer Jim Chalmers gathers a star-studded guest list in Canberra, few in the room will have the expertise to speak with authority on the AI challenge. This is not a minor omission. And I will find it particularly ironic if the 900 submissions have been summarised by Generative AI, instead of having a credible resource in the room. Australia’s problem is not that we lack access to AI tools, it’s that we don’t trust them. Not as individuals, not as consumers, and certainly not in the C-suite. The evidence is stark. A recent KPMG and University of Melbourne study of 48,000 people across 47 countries found Australia ranked dead last in believing AI’s benefits outweigh its risks - just 30 per cent of Australians agreed. Without trust, adoption (and productivity) will stall. Trust rests on two pillars: transparency and data sovereignty. Transparency means knowing what data an AI model was trained on, how it was obtained, and what biases are embedded. That’s not information we will get from a US mega-tech provider - not now, not ever. Data sovereignty means ensuring Australian data stays under Australian control, from storage to processing. Many “onshore” AI services quietly send your information offshore for processing on GPUs located in the US, Brazil or beyond. Most Australians have no idea this is happening. If we want teachers, doctors, lawyers - anyone in critical roles - to embrace AI, they must know the systems they use are trained on Australian-relevant data and their inputs never leave our jurisdiction. This is achievable through sovereign models and federated learning, where data is used for training without ever leaving the state in which it resides. Tesla chair Robyn Denholm is right to warn that we are still in “inning zero” of the AI game and that if we only adopt AI, we might see a $115 billion benefit, but if we build and own the technology, the returns could be ten times greater. That future won’t happen without urgent action now. Yet none of this will reach the Treasurer’s ears if the right voices aren’t in the room. By ignoring AI trust, the roundtable risks squandering the $116 billion productivity prize. If we fail to act now, Australia will fall further behind and no amount of post-summit spin will disguise that failure.

61

25 Comments

🛩️ While everyone was criticising the NZ Government’s national AI strategy, MBIE’s Responsible AI Guidance for Businesses flew under the radar. Despite several shortcomings, the guidance provides a useful introduction to Responsible AI for Kiwi businesses. 👷 But a key challenge is moving from high-level principles and “tips” to practical, proportionate actions that reflect our local environment. Because while the risks of AI are clear (well, not in the guidance unfortunately…), ultimately it’s how you address them that matters. As discussed in our recent Simply Privacy newsletter article (link in comments), we recommend taking action now to embed responsible AI practices. That includes: ✳️ Understanding your AI activities, including shadow AI ✳️ Confirming your AI strategy and risk appetite ✳️ Clarifying accountabilities and responsibilities ✳️ Having clear policies and guardrails around acceptable use of AI ✳️ Training your teams Real progress in driving trust in AI will come not from waiting on government or global standards, but from businesses putting proportionate, workable practices into place. #responsibleai #aigovernance #privacy #trust Simply Privacy Emma Pond Daimhin Warner

39

5 Comments

The Women in Data Science New Zealand Event on August 18th has 17+ Speakers covering the spectrum of Data Science to highlight and attract more women into the field. While Data Science is a relatively new field over the past 10 years, tech in general has been a male dominated field, hence the amazing work of WiDS both globally and locally here is all about shinning a light on the field to make it more attractive to women. From the recently graduated to the super experienced, each diverse story will share a different part of Data Science and the speakers journey. Hosted at Victoria University of Wellington Centre for Science and Artificial Intelligence in Wellington, a long standing partnership with WiDS over it's 8 years, it's going to provide diverse perspectives and great opportunity to connect with women in data science whether you're currently in the industry or looking at entering it! Please register to attend - all are welcome: https://lnkd.in/gcwpz595 Below is the program so you can get the headline of the speakers and subjects: Key Note: - Stop, collaborate and listen - Professor Bryony James Provost at Te Herenga Waka - Victoria University of Wellington Fireside chat: From Stanford University across the oceans to Nelson New Zealand - Dana Briscoe, PhD and Kate Kolich Presentations: - Māori data sovereignty what do data scientists need to know - Ernestynne Walsh - Data Science and AI applications - Junhong (Jennifer) Zhao - Senior Lecturer in Artificial Intelligence - Better understanding pacific strengths with data - Shawn Wimalaratne Social Equity Researcher - Lessons learned in Data and Analytics - Silvina Pugliese (PhD) Think Data Lab Panels: Unlocking Value, Shielding Privacy for AI and Open Data Innovators with: - Liz MacPherson - Deputy Privacy Commissioner - Office of the Privacy Commissioner - Ngapera Riley - CEO Figure NZ - Emma MacDonald - Director Centre for Data Ethics and Innovation Stats NZ - Panel facilitator - Amelia Bentley Data and AI Scientist - HazardCo Navigating Career Journeys in Data Science: Pathways, Pivots, and Purpose - Daria (Ah Reum) Kwon - Manager Integrity Services ACC - Austin McCourt - Data Insights Analyst NZ Police - Jan Sheppard - Chief Data and Analytics Officer ESR - Macushla Howell - Head of Data - NZ Post  Panel facilitator – Jaime Linton Lightening Talks: - Data Science at Scale - from data and research to operational outputs by Poppy Schlaadt and Lydia Acton - All about Social Investment Tahia Eaqub - Interim Director Data and Economics Social Investment Agency In summary, if you or someone you know is interested in Data Science, or doesn't know if they are and would like to explore, this event will give you the insights and the community to be a part of to explore Data Science. Congratulations to the WiDS NZ Ambassadors who make these events happen; Kate Kolich, Ivy Liu, Bing Xue and Qi Chen and thank you to all sponsors.

82

20 Comments

Why ethics? In the FAQ section of the 10th edition of the Australian 2025 Ethics Index I read that: "Ethics is at the base of all decision-making and good governance is decision-making within a sound ethical framework". That's an expectation and a desideratum... or the reality? 🌐 Source: https://lnkd.in/ehYG6Smu 🌐 Past reports (2016-2024): https://shorturl.at/Z0aZu Katrina Horrobin, CEO of Governance Institute of Australia:"One of the standouts of this year’s report is the generational trends, where previously we have seen separate between the generations, we’re now seeing convergence. Gen Z and Baby Boomers, often seen as worlds apart, are now aligned in their ethical outlook... ethical leadership is no longer a niche expectation, it’s a universal one". Again... if data seems saying that, looking at the reality, outside the window, to the announcement, to what's happening in the jobs and activities where AI is replacing rather than supporting humans, are we sure about? The survey included 1.000 weighted respondents representing the Australian adult population (around 27.2 million Australians and ~22 million adults...) by age, gender and location, conducted between July-August 2025 but we must be aware of the potential limitations of the study: it measures opinions rather than actual ethical behavior, it is limited to the Australian citizens, it is limited to a few of months and not a full-year of dynamics and behaviors and of course... respondents may have provided socially acceptable rather than "true" responses. It could happen, we are humans. Launched in 2016 amid royal commissions into banking and aged care, the Index has chronicled ethics evolving from a post-scandal buzzword to a post-pandemic imperative, amplified by COVID solidarity, economic squeezes, now AI anxieties and this insane hype and marketing around the need to be (FOMO, FOSO, ...). The reality unfortunately is different: outside the survey's controlled environment, actual AI deployment often contradicts ethical guidelines and the Corporate AI adoption is mostly focused on efficiency gains rather than ethical implementation with only 1% of companies believing they've reached AI maturity, despite widespread investment. This performance-ethics gap reflects the broader institutional trust crisis the Index documents: people want to see the decision-making process, not just assign blame afterward. If ethics matters as much as the survey suggests, why do ethically questionable practices persist across sectors?  As you can see... something doesn't match. In my opinion the Ethics Index is a valuable diagnostic tool but I believe that treating perception-based metrics as policy rules can lead to governance structures that are based on changing public opinions instead of solid principles. What do you think?   #Ethics #AIEthics #ArtificialIntelligence #AI #Innovation #EthicsIndex #Governance #AIGovernance

28

4 Comments

What I took away from CDAO 2025 Melbourne: 🔹 Governance and agility can co-exist. It’s about delivering incremental value. Think big, start small. 🔹 Data for good is real. Seeing how Foodbank Victoria is using data to improve food distribution and make a real social impact was a powerful reminder that analytics can drive meaningful change beyond just commercial outcomes. 🔹 Strategy needs to be adaptive. In today’s environment, long-term roadmaps are less effective than short-term, responsive planning. 🔹 Empowered accountability and a culture of learning are essential. Innovation thrives when people feel safe to experiment and even fail. 🔹 Cross-functional collaboration requires deliberate intent. It’s not just about structure, but about aligning roles to real impact, especially for technical teams. 🔹 AI and data democratisation is more than just access. It’s about responsible enablement, curated access, and balancing innovation with governance and compliance. 🔹 Human-in-the-loop thinking remains critical. GenAI can do a lot, but it can’t replace creativity, relationship-building, or knowing when to stop. 🔹 AI productivity gains shouldn’t be judged only by commercial output. They should also show up in metrics that reflect improved wellbeing and mental health. 🔹 Don’t build and hope they’ll come. Tie everything to real-world use cases and commercial value. 🔹 Safe spaces for experimentation are vital. Not every piece of data needs to be perfect to be useful. Trust is nuanced. If you attended, what stood out most for you? #CDAO2025 #DataLeadership #InnovationCulture #GenAI #DataDemocratisation #Leadership #Konnexus

44

4 Comments

I am doing a 5-part series on Soowgood's shutdown - This one is my favourite. Part 4: The Customers The part I enjoyed the most was understanding customers - their challenges, pain points, solutions and what they'd be willing to pay for. Here's what I found: Customer #1: Patients Bangladeshis are still learning to trust digital solutions. But for their health? That's another level of trust. Yes, virtual consultations and e-pharmacies exist. But patients don't see them as a replacement for "real" care. When money and health are involved, they want the full experience - the doctor's time, presence, and the physical examination. Out of 75K+ patients served in 18 months, only <5% booked for a real healthcare issue. And if patients won't use it for what matters most, they certainly won't pay for it. That makes predictable growth or profitability nearly impossible. Customer #2: Doctors (& Hospitals) With doctors, conversation is brutally simple: "Can you bring me more patients?" We chose a harder path - awareness and accessibility, not referrals on commission. Ethical, but difficult. Pharmas dominate doctors with incentives. So you either compete or partner with them. We managed the latter, securing sponsorships - proof there was value to unlock. But the bigger challenge was the appointment system (for both doctors and hospitals). On paper, digitising it looked like a huge gap. In reality, it was a maze of manual decision-making. How do you automate the system when: 1. Receptionists are told not to give out serial numbers 1-3, as they are kept for emergencies or VIPs? 2. Odd-numbered serials are given out by receptionists, even numbers by the doctor's secretary? 3. A doctor scheduled to start at 8 pm arrives at 9:30 pm? (Especially when the whole point of automation is to reduce patients' waiting time) How do you automate a system where the rules change by the hour and the exceptions outweigh the standard? The reliance on human discretion and ad-hoc workflows made digitisation nearly impossible. And these learnings are from observing 6 (mid-level) hospital chains, 21 diagnostic centres and working directly with 2 and 180 doctors. In short; - Patients expected free services - Doctors wanted only paying patients - Competitors undercut one another endlessly In the end, no sustainable moat. No pricing power Customer #3: B2B The B2B side looked promising with big checks coming in, but it came with fragility baked in. Partnerships depended more on connections than value. Deals that are signed behind closed doors often struggle to reach the end users. Low awareness meant low adoption, leaders lose interest, creating a cycle of "paper partnerships" Politics played a constant role - a single leadership change could wipe out months of progress. With low barriers to entry and shallow moats, B2B traction rarely lasted. The last part is about how personal milestones and ambitions played a key role in the decision.

41

6 Comments

I had an early morning today, kick starting it with a breakfast hosted by Schneider Electric at the NZ #CIOsummit As with any conference I've been to this year, #AI is the hot topic. Apparently 77% of NZ companies are investing in some form of generative AI, so if you're not using it now you're already falling behind. But what does AI cost? We had some robust conversation at breakfast about the cost of AI, in terms of electricity consumption. Jason Molloy shared his thoughts on the journey to sustainability and the dramatic shifts he's seeing in data centre infrastructure to keep up with the increased demand AI is generating. Who will ensure we put in the extra effort to take the #sustainable path? Is it the responsibility of the government, the private sector, individual consumers, AI itself? It gave me a whole new perspective on the term #ResponsibleAI and what role I can and should play toward decreasing the carbon footprint AI leaves behind.

27

1 Comment