AI in practice: Build trust through understanding

• Associate Professor of AI and Accounting
• Formerly Senior Lecturer and FinTech lead of Master of Business Analytics at Auckland Business School
• Applies cutting-edge AI technologies to solve problems in accounting and finance
• Enjoys engaging conversations, loves to take long walks in nature, avid reader of Tintin and Asterix to learn French

Helen has a deep understanding of the business context and economic forces. Before embarking on an academic career, she worked as an investment banker in Hong Kong. A self-described ‘bit of a nerd’, Helen had been toying with the idea of doing a PhD ever since her MBA studies. “Although the steep learning curve in banking was initially very rewarding and enjoyable, I was also curious about the economic dynamics behind various phenomena, but never had the opportunity to explore them”, she says. Her passion for understanding how things work eventually led her to academia. Helen and her husband Paul Geertsema moved to New Zealand, where they both completed PhDs.

Inside the black box

A recent paper she co-authored with Paul, which also caught our attention, describes a novel machine learning approach to the relative valuation and peer selection of listed companies. In addition to showing the efficacy of contemporary machine learning models in stock valuation, the paper also made the model’s predictions more interpretable, as Helen explains: “We showed that the valuation multiples predicted by the model can be expressed as a weighted average of the valuation multiples of peer companies. These weights can be interpreted as a measure of comparability, with higher weights indicating closely comparable companies or closer peers, and lower or negative weights indicating dissimilar companies or distant peers. This way we can peek inside the black box making the predictions.”

Demystification

Helen’s enthusiasm is palpable. But she recalls how few people could see the point of the exercise when she started the project that led to the paper two years ago. “Many believed that as long as machine learning models predicted well, we didn’t necessarily need to know how they did it.” Two years on, however, explainable AI has become a hot area of research because, as she says, “it is human nature to want to know ‘why’. And it is not just about understanding, but also about trust. Explainability builds trust.”

Explaining AI-assisted decisions to key stakeholders is where her passion lies. As an academic, Helen feels it is her responsibility to demystify complex AI technologies for non-technical audiences. “This”, she believes, “is essential for building trust in AI and speeding up its adoption.”

It's about the people

Helen has taught at New Zealand’s two oldest universities. What attracted her to Vlerick, apart from its reputation as a digital leader in the heart of Europe? “The main reason is the close alignment between my expertise and the School’s strategy. The strong focus on AI and its application in various business areas has attracted high-calibre faculty and researchers. This is an environment where my experience and expertise will be valued, as well as offering space for career aspirations.” Helen likes Vlerick’s collaboration with business, industry and government organisations, as this offers exciting teaching and research opportunities. “But what really sealed the deal is the genuine warmth, consideration and effectiveness of the people at Vlerick - qualities I’ve observed across the board, which is quite impressive.”

Embracing the present, preparing the future

At Vlerick, Helen will supervise and mentor PhD students, as well as design and deliver new courses to meet the growing demand in her field. And she is keen to get started: “Unlike traditional universities, where change can be slow and introducing new courses may take years, Vlerick is agile and responsive. Here, the focus is on identifying emerging growth areas and developing programmes to capitalise on them.”

Asked what she wants to achieve, she replies: “My short-term goal is to continue learning and researching to keep up with the rapid advances in AI. I also want to learn from my colleagues and get better every day. Looking ahead, my long-term plan is to stick with the strategies that have proven effective.” She pauses. “And enjoy the present moment”, she concludes with a smile.