IAS Christopherson Knott Fellows Dr David Chivers (Economics) and Professor John Paul Gosling (Mathematical Sciences) explore Humans, Numbers and now AI
Have you ever wondered why some people seem like walking calculators while others break into a cold sweat at the sight of a spreadsheet? Or why one person trusts an expert's statistics implicitly, while another remains deeply sceptical? These questions were at the heart of the IAS project "Humans and Numbers," which ran over the previous term.
Numbers are unavoidable; they underpin everything from our grocery shopping and health decisions to the news headlines and economic forecasts that we consume. But, as our project confirmed, how we interpret and act upon these numbers is far from a simple calculation. It's a rich tapestry woven from our backgrounds, experiences, psychological responses to uncertainty, and even the way the information is presented. Our proposed questions remained relevant throughout our investigations and discussions: How do we react when experts deploy statistics? Does clarity matter? Does the source influence trust? How does education shape our numerical confidence? And, crucially, how does the rise of AI change the landscape of expertise and data interaction?
Bringing together a diverse team from across the University, the "Humans and Numbers" project moved beyond theory into practical exploration. Rather than just contemplating potential contributions, we actively engaged these questions through targeted activities. We hosted three core workshops tackling pertinent contemporary issues: "The Quantified Self: measuring health and fitness in the digital age" delved into our reactions to constant self-monitoring, with insights led by our colleagues in Anthropology. "Beyond the Stigma: Understanding Vaping Through Numbers and Narratives" brought together researchers and members of the public to examine how data and personal stories shape perceptions of risk. Our final workshop, "Who Counts as an Expert? Numbers, Uncertainty, and the Measurement of Knowledge," directly addressed the challenging interplay between data, expert opinion, and the communication of uncertainty – a theme further explored in our "Stats for Leaders" training event for MBA students.
So, what did we learn? A key insight was the synergy revealed when bringing together perspectives from seemingly disparate fields like economics, mathematics, statistics, and psychology. Understanding how people react and make decisions in the face of numbers and uncertainty requires bridging the gap between quantitative models and cognitive understanding. We observed the varied, often deeply personal, responses individuals have to tracking data and expert pronouncements. Furthermore, our work highlighted the sophisticated guidance already available on harnessing expert knowledge and, critically, communicating the inherent uncertainty often present in data – lessons vital for policymakers and the public alike.
The project catalysed tangible interdisciplinary work. We have begun building a network focused on alcohol use messaging, initiating the creation of a unit-tracking app. Perhaps most significantly, we are developing the "Dealing with Uncertainty" website, a resource designed to provide practical advice on handling and communicating numerical uncertainty for business leaders, policymakers, and the general public – a direct outcome of our collaborative explorations.
In our final seminar of the Epiphany term, we posed a provocative question: In an age of increasingly powerful AI, do we still need humans in the loop for interpreting and acting on numbers? The conclusion was nuanced, but the overwhelming sense was 'yes'. Our project underscored that understanding numbers isn't just about computation; it's deeply intertwined with human psychology, trust, communication, and context. While AI offers powerful tools, the human element – our ability to interpret, question, contextualise, and communicate – remains essential.
"Humans and Numbers" shed light on this vital interaction, revealing the challenges and opportunities in navigating our data-rich world. The insights gained and the work initiated continue to emphasise that understanding the human side of the equation is paramount as we make decisions, develop policy, and shape the world around us, number by number. You can explore some of our findings and resources on the "Dealing with Uncertainty" website and are more than welcome to contact the PIs directly.