Caroline Jay’s engaging talk explored how we use data in decision-making and why human perception often skews our interpretation of it. She highlighted how we systematically underestimate correlations in data visualizations and demonstrated how simple design tweaks—such as adjusting opacity and size—can improve accuracy. Through fascinating eye-tracking studies and real-world examples, she illustrated how attention, cognitive biases, and cultural differences shape our understanding of information. A deep dive into conditional probabilities revealed that even highly trained professionals struggle with interpreting risk, reinforcing the importance of presenting statistical data in clear, intuitive formats. She also shared research showing that people integrate their own prior beliefs when interpreting data, sometimes overriding objective evidence. The talk took a thought-provoking turn with a discussion on AI, where growing trust in automated systems—especially conversational AI like ChatGPT—raises both opportunities and challenges for critical thinking. With enthusiasm, Jay emphasized that while data is a powerful tool, effective communication and an awareness of human behavior are crucial for making informed decisions. Ultimately, she encouraged researchers and policymakers to design tools that truly work for people, ensuring that data not only informs but also empowers decision-making in a meaningful way.

Written with the support of ChatGPT.