Slav Tabachnik on data & AI in 2024 and beyond

"Many clients I work with assume that a proof of concept (POC) can seamlessly transition into production and yield immediate growth. The reality is much more complex!"

I had the pleasure of working with Slav many years ago when we were both at Deloitte. I’ve always found his way of looking at the world inspiring so we’ve kept in touch. After becoming a Deloitte Partner in 2017, a bit over a year ago Slav stepped into the role of National Generative AI Transformation Leader in the Deloitte Australia Transformation Office. His remit includes “enabling resilient and innovative transformation” across the consulting practice, anticipating and harnessing Gen AI based disruption before it comes knocking.

Slav, 2024 has been another busy year for data and AI. What’s one development and milestone/news story that really caught your eye?

2024 has indeed been a whirlwind of developments in data and AI. While there’s been a flood of innovations, one standout for me is the application of large language models in translation, particularly for interspecies communication.

For example, breakthroughs in decoding the complex communication patterns of dolphins and whales or the vocalisations of elephants have pushed the boundaries of AI's capabilities. These advances not only demonstrate the sophistication of modern AI but also invite us to reconsider our place in the natural world and deepen our understanding of non-human intelligence.

You’ve been working in and around data and AI for a while now. Many things have changed, but tell us about something that was true when you started out in the space and is still important today.

Yes, the landscape of data and AI has changed dramatically, but some fundamentals remain constant. When I first started in this space over 20 years ago, understanding data and ensuring its quality were critical—and that hasn’t changed. Models, no matter how advanced, are only as good as the data they’re trained on. Equally enduring is the importance of data storytelling. Transforming technical insights into actionable business value is what drives decision-making and real-world impact. These principles continue to underpin success in an era of rapid innovation.

It’s been a heady couple of years, with 2024 almost as frothy as 2023. What’s one common misconception about AI that you wish would go away?

A common misconception about AI that I wish would disappear is the belief that AI can ‘understand’ or ‘reason’ like humans. This misperception often leads to misplaced trust in its abilities, as well as unrealistic expectations. For instance, many clients I work with assume that a proof of concept (POC) can seamlessly transition into production and yield immediate growth. The reality is much more complex: implementing AI solutions requires rigorous testing, design, and strategic planning, especially with the increasing scrutiny from regulators. We need to demystify AI’s capabilities while appreciating the discipline and effort required to realise its potential.

The festive season is almost upon us, so many readers will have a bit of extra time to read, learn, reflect. Who do you follow to stay up to date with what’s changing in the world of data and AI?

I read voraciously and struggle to single out just one resource, but I highly recommend tools that make staying informed easier. For instance, Google’s NotebookLM has been transformative for me. It converts dense research papers, like those from arXiv, into podcasts, making it easier to digest cutting-edge AI developments on the go. It’s particularly useful for breaking down technical content into actionable insights. I also explore offerings from ElevenLabs and other similar companies in this space. Beyond tools, following thought leaders like Andrew Ng and organizations like OpenAI, DeepMind, DataBricks and others helps me stay attuned to the latest trends.

Leaning into your dystopian side for a moment, what’s your biggest fear for/with/from AI in 2025?

My biggest fear for AI in 2025 is its escalating use in warfare. Growing up in Ukraine, I’ve watched the conflict evolve with alarming reliance on AI-driven technologies. For instance, Ukraine now produces millions of drones annually, some are now equipped with image recognition and autonomous targeting and target lock on capabilities. While humans remain part of the decision loop, this trajectory raises significant ethical concerns. Allowing machines to make life-and-death decisions could erode accountability and introduce catastrophic risks. Aside from ending this insane conflict, it’s imperative that we establish clear international frameworks to regulate the deployment of AI in warfare before it’s too late.

And now, channeling your inner optimist, what’s one thing you hope to see for/with/from AI in 2025?

I’m optimistic about AI’s transformative potential in personalised medicine and human health in 2025. Inspired by Ray Kurzweil’s vision in The Singularity Is Nearer, I’m particularly excited by advancements like DeepMind’s protein-folding breakthroughs, which could unlock personalized treatments and even eradicate diseases. Imagine a world where AI helps tailor cancer treatments to individual patients or accelerates the discovery of vaccines. With the combined efforts of visionary researchers and innovative companies, I truly believe AI can deliver breakthroughs that enhance lives and redefine the future of healthcare.