Data Visualization and AI: transforming insights into action

How AI-powered data visualization is reshaping education, design, and decision-making — a closer look at its transformative potential with Paolo Ciuccarelli.

Artificial intelligence is revolutionizing the way we approach data visualization, turning complex information into actionable insights like never before.

From enhancing learning experiences in education to driving innovation in design, AI is unlocking new possibilities across industries.

Today we’re surfing with Paolo Ciuccarelli, Architect and Communication Designer, Partner of The Visual Agency.

How do you see the role of AI in enhancing data visualization, and what are its biggest threats and opportunities?

While I remain mindful of AI’s limitations and potential risks, I approach these technologies with cautious optimism.

AI is already proving valuable in many phases leading up to the creation of a visualization. It can identify patterns, enrich datasets, assist in exploratory analysis, and even support adaptive narratives that tailor data representation to specific users or contexts.

However, when it comes to the form-giving moment, deciding how data should be visually expressed, AI still lacks the capacity to move beyond standardized templates or catalogued visual conventions. It is especially limited when visualization is used for rhetorical purposes or when innovation is required in the visual language itself.

In this respect, human designers remain essential.

In certain cases, AI might even bypass visualization entirely by providing direct insight. When that occurs, it is often because the insight is sufficiently simple or self-evident that a visual representation would not have added significant value in the first place.

What challenges do you face when creating interactive designs that effectively communicate complex data to diverse audiences?

One of the central challenges is clarifying that different users often require different visualizations. The idea that a single solution can serve all is persistent but misleading. This misconception stems from a broader misunderstanding: people often perceive visualization as a final product rather than a process.

In truth, the effectiveness of a visualization lies not in what is shown, but in how it engages the user’s cognitive and perceptual system. I often summarize this idea with the phrase: What You Get Is Not (Only) What You See (WYGINOWYS).

Insight does not reside in the visualization alone but emerges in the interaction between the image and the viewer’s mind.

This means the user plays an essential, active role and must be seen as a co-producer of meaning. With complex data, this becomes even more critical, as variations in users’ ability to navigate complexity can significantly influence the insights outcomes.

Can you share an example where interactive design has significantly enhanced understanding or engagement with data?

Much of my work is driven by a mission to make complexity more accessible, an effort in which interactivity plays a vital role. As datasets grow in scale and relational depth, traditional static visualizations often fall short, particularly in fields like network science, where graphs may contain hundreds of nodes and thousands of connections.

Even experts can struggle to interpret such dense static representations. In these contexts, interactive and dynamic visualizations become essential, enabling users to progressively unfold information along paths shaped by both the system's structure and the user’s intent.

The most effective visual experiences combine a balance between author-driven guidance and user-driven exploration. Without such hybrid strategies, we risk falling back on overly simplistic portrayals, an approach that can lead to poor decisions and unintended consequences at both organizational and societal levels.

How to balance aesthetic design with the need for clarity and functionality in data-driven projects?

This is a recurring misunderstanding: that aesthetics and function are somehow at odds. In reality, every visual element carries an aesthetic dimension, whether it is intentionally designed or not. Even the simplest bar chart conveys a visual rhetoric, often one of objectivity, rigor, or neutrality.

These qualities may be desirable in some contexts, but misleading in others.

Many of the phenomena we study and aim to influence, particularly in emerging domains such as AI or quantum computing, are complex, imprecise, and ambiguous. Yet our visual languages often lack the capacity to express such qualities with appropriate nuance.

The challenge is not to eliminate aesthetics, but to integrate them more consciously into our functional goals. This is an area still ripe for exploration.

In your opinion, how can technology be used to make data more accessible and impactful in fields like education or information design?

Focusing on education, I believe we face an urgent need to improve data literacy and empower individuals to actively engage with data-driven discourse. Data informs a growing number of decisions, from the personal to the geopolitical. Yet the ability to participate meaningfully in these conversations remains limited.

While many efforts have focused on improving comprehension—making data more accessible and interpretable—this alone is insufficient. We must go further: enabling people not only to understand data but to transform it, to contribute their interpretations back into the dialogue, and to influence outcomes.

At present, we are producing more informed readers, but not enough active contributors. Meeting this challenge requires an entirely new set of tools, infrastructures, and practices—none of which exist yet in a fully developed form.

Designing for that space is, I believe, a crucial opportunity for the future of information design. And we are ready for the challenge.

Paolo Ciuccarelli is an architect and communication designer known for his work in information design and data visualization. He is Professor of Design and Director of the Center for Design at Northeastern University, Boston. Previously, he spent two decades at Politecnico di Milano, where he led the Communication Design program and founded the award-winning DensityDesign Research Lab.

His research transforms data into meaningful tools and narratives for decision-making, particularly in complex social contexts involving non-experts. He explores the rhetorical and poetic dimensions of design to enhance public engagement and investigates meta-design in relation to emerging technologies like artificial intelligence.

Ciuccarelli has received international recognition, including best paper awards and speaking invitations at Eyeo, Congreso Futuro, TEDx, and institutions such as the Royal College of Art, ENSCI Les Ateliers, Stanford Humanities Center, MIT Media Lab, and Harvard GSD. He is a Senior Affiliate at Harvard’s metaLAB and co-editor of Big Data & Society. His recent projects include a collaboration with Lexus for Milan Design Week 2025 and contributions to the Biennale of Architecture in Shenzhen (2022) and Venice (2025).