Participation at CSCW 2026 with two contributions on Attention-Capture Damaging Patterns

Published on October 14, 2026

This week, I was at CSCW 2026 (ACM Conference on Computer-Supported Cooperative Work and Social Computing) in Salt Lake City, Utah, USA, where I presented two works: a full paper published in the Proceedings of the ACM on Human-Computer Interaction (PACM HCI) and a poster.

Both contributions tackle the same problem — Attention-Capture Damaging Patterns (ACDPs) such as infinite scroll, autoplay, and casino-like pull-to-refresh — from two complementary angles: the designers who build digital interfaces, and the users who experience them every day.

Full Paper: Digital Wellbeing by Design

On Tuesday, October 13, I presented our paper “Digital Wellbeing by Design: Safeguarding End-Users’ Time and Attention in UI/UX Practices” in the Ethics and Responsible AI session. The paper was co-authored with Luigi De Russis and Massimiliano Pellegrino from Politecnico di Torino.

Traditionally, digital wellbeing has been considered an end-user responsibility: digital self-control tools like timers and usage dashboards ask people to fight, on their own, against interfaces engineered to capture their attention. In this work, we flipped the perspective and asked how we can raise designers’ awareness of ACDPs — and surface credible alternatives — early in the design process and inside the tools they already use.

To this end, we developed Digital Wellbeing Lens, a plugin for Figma that acts as a “digital-wellbeing assistant.” Leveraging a multi-modal chain of prompts to a Large Language Model, the plugin scans UI mockups, highlights potential ACDPs and their risks for end users, and suggests alternative designs that preserve similar functionality (e.g., replacing infinite scroll with pagination). Our work unfolded in three phases:

  1. Prompt development and evaluation: we developed a detection prompt, tested it on a dataset of 10 Figma prototypes recreated from examples in the HCI literature, and asked 4 recognized experts in deceptive design to assess its detections and alternatives
  2. Refinement and implementation: building on the experts’ feedback, we refined the technique into a multi-modal chain of prompts that combines images and JSON representations of the UI, and integrated it into the Figma plugin
  3. Evaluation with designers: we conducted a formative study with 15 professional UI/UX designers, using the plugin as a design probe

Designers found the plugin’s feedback generally helpful: 54% of the detections were rated as very helpful, about two-thirds of the suggested alternatives were considered helpful or very helpful, and around 60% of them were expected to improve usability. At the same time, participants pointed out that engagement KPIs often sit above designers’ choices, and that implementing some alternatives may conflict with company goals. Our results suggest that the primary value of such a tool lies in sparking organizational discussion about design trade-offs rather than prescribing specific design changes: the plugin can work as a boundary object that designers bring into conversations with product managers — ideally complemented by regulation.

Below you can browse the slides from my presentation:

Loading...

📥 Download PDF

Poster: ACDPs Transparency Labels

On Monday, October 12, during the Poster and Demo Reception, I presented the poster “ACDPs Transparency Labels: Supporting User Agency Against Attention-Capture Damaging Patterns,” co-authored with Riccardo Graziosi and Luca Passarella.

The poster introduces ACDPs Transparency Labels (ATLs): contextual, in-app interface notices that make ACDPs visible, explain how they may affect users’ digital wellbeing, and offer a clear call to action to regain control. For example, an ATL on YouTube can warn users that autoplay is active, show how many videos have played in a row, and offer a one-tap toggle to disable the feature. Inspired by transparency mechanisms such as cookie notices and nutrition labels, ATLs extend the concept beyond data collection to interface-level persuasion.

We developed the concept through two co-design workshops with 13 users of TikTok and YouTube, and distilled the most generative findings into five preliminary design considerations:

  1. Visual identity: ATLs should respect the platform’s look and feel, including typography, color palette, and interaction patterns
  2. Language: ATLs should adopt a concise, jargon-free, and non-judgmental tone that supports reflection without blaming users
  3. Timing and position: ATLs should appear when an ACDP is activated, close to the interface element they refer to
  4. Adaptability: ATLs should adapt their tone and visual design over time to counter habituation
  5. Interactivity: ATLs should offer an immediate way to act on the ACDP, e.g., disabling autoplay or switching to a non-profiled feed

Rather than forcing platforms to abandon engagement-based features, ATLs make them more visible, understandable, and accountable to users. In doing so, they may help platforms build trust and give regulators a concrete, interface-level mechanism to operationalize the transparency obligations of the EU Digital Services Act and AI Act.

It was great to be at CSCW and to discuss these works with the community. Many thanks to all my co-authors, and to everyone who stopped by the poster or joined the discussion after the talk!

More Information