本文探讨了如何运用电影手法增强数据叙事的吸引力。文章回应了关于冲突、蒙太奇和AI在数据叙事中的作用的讨论。作者强调了冲突在故事中的核心地位，介绍了蒙太奇技术在控制叙事节奏和增强观众参与度方面的作用，并分析了AI在数据叙事中的多种角色，包括创作、优化、审核和辅助。通过这些方法，数据叙事可以变得更具吸引力和影响力。

🤔 数据叙事的核心在于冲突。文章指出，即使故事看起来没有冲突，电影和数据叙事都离不开冲突。冲突可以是外部的，也可以是内部的，但它驱动了故事的叙事张力。文章强调，如果数据没有揭示问题、异常或紧张点，那么就应该采用数据呈现模式，而不是强行构建故事。

🎬 蒙太奇技术增强了数据叙事的表现力。文章详细介绍了蒙太奇技巧，例如闪回和并行叙事，它们可以用来控制数据叙事的节奏和焦点，进而提升观众的参与度。通过蒙太奇，数据故事可以从纯粹的逻辑结构转变为同时考虑人类注意力、情感和记忆的结构，从而使故事更具吸引力。

🤖 AI在数据叙事中扮演着多重角色。文章深入探讨了AI在数据叙事中的应用，包括创作、优化、审核和辅助。AI可以快速生成文本、总结数据集，甚至创建视觉元素。然而，作者强调，AI应该作为人类故事讲述者的助手，而不是替代者。同时，文章也提醒了在使用AI时需要注意伦理问题，避免偏见和误导。

Data storytelling involves communicating the results of data analysis through a compelling narrative. In a previous post,¹ I examined how cinematic techniques, including narrative arcs, character roles, and visual framing, can be applied to data storytelling. That discussion focused on foundational concepts, emphasizing how a three-act narrative structure can help organize insights into a compelling journey anchored in the data itself. 

During a recent talk² on the same topic, a wide range of questions emerged from the audience. These questions explored ethical boundaries, narrative diversity, expert audiences, and the expanding role of AI in data storytelling. Some asked whether data stories must follow conflict-driven models, and others wondered how montage could be used to tell richer, more layered stories. And many were curious about how AI tools might shift the very process of crafting a story: Will they augment creativity, or automate it?

This post follows up on those questions. It directly addresses the themes that emerged during the discussion, organizing them around a few key axes: emotional tension and ethics, structural alternatives such as montage, adapting cinematic storytelling for technical audiences, and the role of AI as both a tool and a collaborator. Each section responds to the community’s curiosity, using their questions as a lens to explore what cinematic data storytelling might become next.

Do We Always Need Conflict?

One insightful question from the audience was:

“There does not always need to be a conflict in storytelling, right? For example, a story could be told about the journey of the inferences from the data that has not been told before. Even in movies, not all the movies fit into the three-act structure. As long as it is new, interesting, and engaging, that’s what matters. What are your comments on that?”

Even when a story appears to lack conflict, even in cinema, this is rarely the case. In fact, every story requires some form of conflict. In film, this typically falls into two categories: external conflict, where the protagonist faces an obstacle outside of themselves, and internal conflict, where the struggle is psychological or emotional in nature. However, regardless of the type, every story involves some kind of transformation, namely, a journey that takes the protagonist from one state to another. This principle holds even when a story doesn’t follow a classic three-act structure. There is always tension to resolve, whether it’s visible or subtle, external or internal. Without conflict, there is no story, just a sequence of events or facts.

The same applies to data storytelling. If the data doesn’t reveal a problem, anomaly, or point of tension, then what we’re doing isn’t storytelling, it’s data presentation. More broadly, as shown in Figure 1, there are three main modes of data communication:

Data Reporting: Sharing all the data, often without interpretation.Data Presentation: Select and present relevant insights, often through structured slides.Data Storytelling: Selecting one compelling insight and developing it through a narrative.

Data storytelling is not just a subset of reporting or a more polished presentation, but a method that involves narrative construction. And narrative, by its nature, depends on conflict. If the data reveals no challenges, no surprises, and no obstacles, then it may be best served by the presentation model rather than a forced story arc.

This doesn’t make it less valuable. It just means it belongs to a different mode of communication. As storytellers, we must be honest about this distinction. We shouldn’t stretch insight into narrative where it doesn’t exist. However, when conflict is present, it is then that the data reveals change, risk, inequality, deviation, or potential, and a story can emerge, complete with a hero, a problem, and a path toward resolution.

Montage Techniques

Another interesting question from the audience was:

“From a cinematic standpoint, do you build in a ‘money shot’? Example: the camera zooming out and up from Gary Cooper in High Noon, where you see that he is all alone and no one has come to help him. If so, how might you do so, perhaps using ChatGPT to generate as you mentioned earlier in your presentation?”

The short answer is yes, but it’s not only about what we show. It’s about how we sequence it. This is where montage techniques come into play. Montage doesn’t alter the story itself; it reshapes how that story unfolds. It’s the difference between linear narration and editorial design, and it offers data storytellers powerful tools for adapting their stories to specific audiences.

In traditional data presentations, particularly those aimed at executives or decision makers, there is often pressure to lead with a summary: the insight, the number, the decision point. The idea is to respect time. But this approach can backfire. Once the main point is revealed, attention drops. The rest of the story is reduced to justification, not discovery.

Here, montage offers a valuable alternative.³ Consider the use of a flashback, a cinematic device that shows a critical moment upfront, then rewinds to explain how we got there. Applied to data storytelling, this means beginning not with the conclusion, but with the moment just before the insight emerged. A surprising shift in a graph. A troubling anomaly. A decision point where something doesn’t add up. This opening builds tension and encourages the audience, even the busiest executives, to follow the thread to its resolution.

Montage can also support parallel narratives, useful when comparing two divergent paths, populations, or hypotheses. Or it can apply rhythmic editing, where alternating patterns of visual and verbal pacing keep audiences engaged over time.

Another critical concept here is the hook, the narrative device that grabs the audience’s attention at the start. In cinema, the hook is the opening image, the unresolved question, the promise of change. In data storytelling, the hook must be deeply audience-specific. What matters to a public audience (e.g., societal implications) might not matter to a technical team (e.g., model accuracy), or a business leader (e.g., financial impact). The hook is what makes them care, and it needs to arrive early, with precision.

Ultimately, montage techniques help control the tempo and focus of a data narrative. They enable us to transition from a purely logical structure to one that also considers human attention, emotion, and memory. Used ethically and intentionally, they turn stories into experiences even in a data-driven world.

The Role of AI in Data Storytelling

Another question that stood out was the following:

“What’s the future of this field? Will AI automate cinematic data storytelling?”

As the data storytelling landscape evolves, one of the most transformative forces reshaping it is generative AI. We are no longer in a phase where AI simply automates routine tasks; it is becoming a collaborative partner in the creation, refinement, and delivery of data-driven narratives. 

According to Li et al.,⁴ AI can play four distinct roles in data storytelling:

Creator – AI can generate first drafts of texts, summaries of datasets, or even visual elements like infographics. Tools like ChatGPT and DALL·E can produce narrative or visual scaffolding in seconds. However, outputs in this mode often lack depth or originality unless carefully guided. The risk is homogeneity, where stories that appear different in data share the same tone.Optimizer – AI can refine existing content, improving readability, adjusting tone, or restructuring material for better flow. This is especially helpful when a story needs to be tailored to different audiences. For example, it can transform a technical explanation into something digestible for a non-expert, or make a summary more persuasive for an executive audience.Reviewer – AI can act as a quality control mechanism. It can identify inconsistencies in logic, flag vague sections, or point out misalignments between visuals and text. While it can’t replace a human editor, it enhances the revision process and accelerates iteration.Assistant – This is perhaps the most potent and versatile role. As an assistant, AI supports tasks such as data collection, document summarization, generating alternative plot structures, translating content, and creating audience-specific versions of a story. For example, it can suggest new “hooks” depending on whether you’re speaking to scientists, policymakers, or the general public.

What’s crucial is that AI does not replace the storyteller. It amplifies their voice, allowing for greater creative range and faster execution, but only when the human remains in control. 

Equally important are ethical considerations. AI can introduce biases, hallucinate content, or frame data in misleading ways.⁵ To mitigate this, storytellers should apply robust validation methods, such as the Retrieval-Augmented Generation (RAG) technique,⁶ and continually review outputs for accuracy, completeness, and fairness.

Conclusion

This article builds on a previous post that explored cinematic techniques in data storytelling, responding to questions raised by readers and participants during a recent talk. These questions opened up deeper reflection on the ethical, structural, and technological dimensions of narrative design in the context of data.

The discussion began with an examination of the fundamental role of conflict. Every story needs a conflict, either external or internal. If you can’t find a conflict, don’t tell a story. Focus, instead, on data presentation.

Shifting to montage techniques, we highlighted how story sequencing affects engagement. Especially for executive audiences, strategies such as flashback or parallel editing can help preserve attention while maintaining a logical structure. Techniques borrowed from cinema become tools for audience-centered design.

Finally, we examined the growing role of AI, not as a replacement for human insight, but as a collaborator in crafting, refining, and tailoring stories. AI can help scale narrative strategies to different contexts and audiences, provided it is guided critically and used ethically.

References

1. Lo Duca, A. (2025). The Narrative Power of Data. Retrieved from https://cacm.acm.org/blogcacm/the-narrative-power-of-data/ (Last Access 2025/06/08)

2. Lo Duca, A. (2025). Applying Cinematic Techniques to Data Storytelling. Retrieved from https://events.zoom.us/ev/AiRD4Q7FOZsekO5YTvyw5_0c_7KBlMFPHjKpslAjIPCDTupwue9T~AgZiB7bIfC0-tX2iDnaaRTHbA5zdO0QYqC2UJgKE-EA82oNq39mXX94KUDTRzT0tCxO-tbYj3VPY3kYsHB53Y8rYmw (Last Access 2025/06/08)

3. Lo Duca, A. (2025). Become a Great Data Storyteller. John Wiley & Sons.

4. Li, H., Wang, Y., & Qu, H. (2024, May). Where are we so far? Understanding data storytelling tools from the perspective of human-AI collaboration. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems (pp. 1-19).

5. Floridi, L. (2023). The ethics of artificial intelligence: Principles, challenges, and opportunities. Oxford University Press.

6. Lewis, P. et al. (2020). Retrieval-augmented generation for knowledge-intensive NLP tasks. Advances in Neural Information Processing Systems, 33, 9459-9474. 

Angelica Lo Duca is a researcher at the Institute of Informatics 
and Telematics of the National Research Council, Italy. Her research  interests include data storytelling and the application of AI to 
different domains, including cultural heritage, tourism, education, and more. She is the author of Data Storytelling with Altair and AI (Manning Publications, 2024) and Become a Great Data Storyteller (Wiley, 2025).