Neural Notes: AI trends and challenges to watch in 2025

Welcome back to Neural Notes, a weekly column where I look at some of the most interesting AI news in Australia. In this edition, industry experts weigh in on their AI predictions for 2025.

Aurélie Jacquet: Standards, certification and international collaboration

Aurélie Jacquet, director of Ethical AI Consulting and chair of Standards Australia’s AI Committee, shared insights into the priorities for AI standards and governance in 2025.

She highlighted the growing emphasis on certification frameworks and their role in shaping the ecosystem.

“From the standards perspective, I think the focus is on certification, most definitely – how to support the certification of organisations for the use of AI systems,” Jacquet said.

“So that’s our main focus for 2025 – to support the ecosystem.”

She then turned her attention to international developments, such as data spaces and AI factories, and their potential implications for Australia’s AI infrastructure.

Jacquet explained data spaces are designed to enable secure collaboration and data exchange, with governance integrated from the outset.

“It’s [about] looking at what infrastructure we need in Australia to promote and develop AI well, so a focus on infrastructure.

“In the EU, there’s been a lot of attention on the EU AI Act, but somehow a lot less attention about what they do to also consider how to do AI well.

“Data spaces are creating an exchange… where you can trade and collaborate, but with the governance spec team ready.

“There’s been a sense of this data exchange with governance and the architectures baked in at the forefront.”

Jacquet also discussed the changing realities for organisations as they scale AI use cases, stressing the need for updated processes and collaboration.

“It’s how the day-to-day practice will change. The reality is going to hit next year more so, because you realise that you need to work a lot closer together to manage AI, have a much more interdisciplinary practice.

“And when you have some things that change regularly, you need to also start ongoing compliance and monitoring,” Jacquet said.

Dr Ariella Heffernan-Marks, Ovum AI: AI for closing the gender data gap in health

Dr Ariella Heffernan-Marks, founder and CEO of Ovum AI, envisions significant advancements in gender-specific AI by 2025.

“We often see women’s health and safety being reduced to talking points and political arguments. My hope is that having robust data around women’s health will validate women’s experiences and lead to meaningful conversations and change. A female-focused AI can help drive this change,” Dr Heffernan-Marks said.

She highlighted the intersection of AI and femtech as a powerful force for improving health outcomes.

“There is still an enormous gap in women’s health data and its use in AI models, which are gendered and often misrepresentative of women. AI chatbots have been shown to effectively help with improving healthy behavior change, from reducing smoking to increasing medication adherence.”

Heffernan-Marks believes venture capital investment in Australian-created, gender-specific AI will play a pivotal role in driving innovation, bridging the health data gap, and fostering partnerships across the health and tech sectors.

“The safe and ethical use of artificial intelligence can address the gaps in health data, health literacy, and motivation, and I’m optimistic about the community of leaders in this sector driving progress,” Dr Heffernan-Marks said.

Robin Marchant, Klaviyo: Hyper-personalised customer experiences

Robin Marchant, senior marketing director APAC at Klaviyo, predicts transformative shifts in customer experiences through AI-powered conversational commerce.

“AI-powered conversational commerce will continue to take off, with virtual assistants evolving into intuitive shopping guides. These chatbots won’t just answer questions; they’ll actively support customers as they navigate their shopping journey, creating an interactive experience that feels personal and engaging.”

He further anticipates hyper-personalisation becoming the new standard.

“Imagine personalisation going beyond just recommendations, where AI dynamically adjusts website layouts, messaging, and timing for each shopper, crafting a genuinely unique experience,” Marchant said.

“This hyper-personalisation will deepen customer connections and boost conversions like never before for brands.”

Julian Stevenson, RMIT Online: Smarter AI implementation

Julian Stevenson, product and operations director at RMIT Online, reflected on how organisations will use AI more strategically in 2025.

“The last couple of years have seen companies grapple with the impacts of AI, distracted by the potential of AI to reduce workforces or make roles redundant,” Stevenson said.

“As the dust continues to settle, organisations will become smarter about defining specific opportunities for removing mundane tasks and augmenting roles rather than focusing on replacing roles.”

Stevenson also pointed to the lessons learned from AI’s flaws and limitations.

“The proliferation of AI flaws — like seven-fingered hands and biased text — has helped balance the discourse. Businesses are learning that AI isn’t flawless and needs to be implemented with care, forcing organisations to be more focused in selecting both the job task and the AI tool to solve specific business problems.”

Kristen Migliorini, Komply AI: Moving towards predictive AI governance

Kristen Migliorini, founder and CEO of KomplyAi, outlined three critical shifts needed in Australia’s AI landscape by 2025.

“First, a massive investment in upskilling and AI literacy across all sectors – not just tech,” Migliorini said.

“Second, we need to fundamentally reimagine how we collaborate – government, large industry, innovators – leaning on our collective intelligence, both human and machine, sharing resources, skills, and agile innovative thinking in ways we’re currently failing to do. 

“Third, a move toward predictive rather than reactive AI governance.”

Migliorini highlighted the importance of shifting the regulatory conversation from loss aversion to value creation.

“Our focus at KomplyAi is on moving AI governance, risk, and compliance from risk identification to risk prediction,” Migliorini

“What I call the ‘fundamental safety shift’ is needed before organisations can safely scale AI into customer-facing and higher-risk areas. Is that fighting AI with AI fire? Yes, but that’s what is needed to keep pace.”

Danny Wu, Canva: Generative AI and proactive customer experiences

Danny Wu, head of AI products at Canva, foresees personalisation reshaping the customer experience in 2025.

“The early days of generative AI have largely been one-size-fits-all solutions. While AI adapts to the context, the vision of generative AI that learns from each customer interaction is something that will become commonplace,” Wu said.

Wu highlighted the potential of AI to move beyond reactive service, saying, 

“AI will be used to proactively predict customer preferences, offering recommendations and solutions before issues even arise. This shift will redefine customer loyalty as consumers gravitate towards brands that can anticipate their needs with precision.”

Elena Tsalanidis, Deeligence: Rethinking generative AI for legal work

Elena Tsalanidis, co-founder of Deeligence, predicts the current chat-based generative AI models will be replaced by more sophisticated solutions.

“No more fumbling around to get what you need. Vendors with great apps do all the RAG/prompting so you don’t have to,” Tsalanidis said.

That’s the value we bring as a vendor to the highly skilled, judgment-based work done by corporate lawyers. They don’t have time to finesse prompts—they need legal work drafted by AI delivered to a high quality right out of the gate.”

Never miss a story: sign up to SmartCompany’s free daily newsletter and find our best stories on LinkedIn.