[Paper Review] Visualization for Villainy

Andrew McNutt, Lilian Huang+1 more|arXiv (Cornell University)|Sep 13, 2021

Data Visualization and Analytics49 references4 citations

TL;DR

This paper introduces a design space for villainous visualization, reframing visualization techniques to intentionally cause harm by exploiting deceptive design, environmental destruction, and data misuse. It unifies existing deceptive practices with novel malicious tactics—such as emotional spear phishing and data physicalization for propaganda—advocating for visualization as a tool of systemic harm rather than ethical analysis.

ABSTRACT

Visualization has long been seen as a dependable and trustworthy tool for carrying out analysis and communication tasks -- a view reinforced by the growing interest in applying it to socially positive ends. However, despite the benign light in which visualization is usually perceived, it carries the potential to do harm to people, places, concepts, and things. In this paper, we capitalize on this negative potential to serve an underrepresented (but technologically engaged) group: villains. To achieve these ends, we introduce a design space for this type of graphical violence, which allows us to unify prior work on deceptive visualization with novel data-driven dastardly deeds, such as emotional spear phishing and unsafe data physicalization. By charting this vile charting landscape, we open new doors to collaboration with terrifying domain experts, and hopefully, make the world just a bit worse.

Motivation & Objective

• To reframe visualization not as a tool for good, but as a deliberate instrument of harm, countering the dominant narrative of visualization as inherently trustworthy.
• To address the lack of scholarly attention to visualization's potential for intentional, malicious use, especially in contrast to its growing role in social good.
• To create a structured design space that categorizes and systematizes methods of visualization-based harm, enabling deeper study and collaboration with 'villainous' practitioners.
• To challenge the status quo by highlighting how mundane visualization practices can perpetuate structural harm when left unexamined.
• To propose new malicious applications—such as climate data misrepresentation and data physicalization on natural landscapes—demonstrating how visualization can amplify systemic inequities.

Proposed method

• Develop a two-dimensional design space for villainous visualization, partitioning harm by type (physical vs. non-physical) and impact (direct vs. indirect), inspired by the Hippocratic Oath’s 'first do no harm'.
• Reframe the ethical imperative as 'first, let us do harm', using this inversion to structure a taxonomy of malicious visualization techniques.
• Integrate and extend prior work on deceptive visualization (e.g., Snider, Correll, Tomlinson) into a unified framework for intentional harm in visualization.
• Introduce novel malicious tactics such as emotional spear phishing via visual cues, climate data manipulation to exacerbate inequality, and data physicalization on natural landscapes (e.g., carving logos into mountains or rainforests).
• Propose the use of energy-intensive data warehousing and environmental degradation as forms of 'autographic visualization' reflecting capitalist accelerationism.
• Advocate for curating harmful example datasets (e.g., preserving eugenicist datasets like Fisher’s iris data) and enumerating 'dark charting patterns' to systematize malicious design.

Experimental results

Research questions

• RQ1 How can visualization be intentionally designed to cause maximal harm, rather than serving ethical or analytical goals?
• RQ2 What structural and design-level mechanisms enable visualization to perpetuate systemic harm, even when not overtly malicious?
• RQ3 In what ways can data physicalization and environmental destruction be leveraged as forms of visualization-based villainy?
• RQ4 How might the normalization of data collection and visualization practices contribute to long-term, irreversible harm, such as climate injustice?
• RQ5 What role can visualization researchers play in enabling or mitigating the misuse of visualization tools for malevolent ends?

Key findings

• The design space successfully unifies prior work on deceptive visualization with novel malicious applications, such as emotional spear phishing and data physicalization on natural landscapes.
• The paper identifies that many current visualization practices—while not overtly evil—nonetheless reinforce dominant power structures and perpetuate harm through inaction or poor design choices.
• Environmental destruction, such as deforestation for visual art or data center energy consumption, is framed as a form of 'autographic visualization' reflecting capitalist accelerationism.
• The authors argue that the most effective evil is not through elaborate schemes, but through maintaining the status quo: normalizing data as factual, ignoring ethical implications, and avoiding critical reflection.
• The paper critiques the ethical imperative of 'do no harm' in visualization, proposing its inversion as a necessary step to understand and counteract the real-world harms visualization can enable.
• By advocating for the curation of harmful example datasets and the enumeration of 'dark charting patterns', the paper provides a framework for understanding and potentially mitigating malicious visualization practices.