paper / peterchapman / Mar 13
As social media litigation matures across U.S. jurisdictions, this paper proposes a concrete, implementable abatement mechanism applicable to settlements or court-ordered remediation — going beyond financial compensation toward structural reform. Drawing on American product safety litigation precedent, the authors define requirements for a harm-mitigation procedure that sits at the intersection of legal process, public health assessment standards, and modern platform engineering constraints. The framework explicitly addresses privacy tradeoffs and oversight implications, positioning it as a practical policy instrument rather than a theoretical exercise.
social-mediaplatform-safetypublic-healthcontent-moderationtech-regulationlitigationharm-abatement
“Existing social media harm lawsuits exist within a established tradition of American product safety litigation, making non-monetary remediation legally precedented.”
paper / peterchapman / Jun 19
Hoop Diagrams offer a circular visualization for set data, representing sets as hoops and intersections as sectors. A user study comparing them to Linear Diagrams showed comparable usability, with Hoop Diagrams excelling in accuracy for certain questions and maintaining a consistent square aspect ratio, unlike Linear Diagrams which can expand horizontally. The choice of diagram type may depend on the presentation context.
data-visualizationset-theorydiagramsuser-interfacesinformation-designhmi-design
“Hoop Diagrams are a new circular visualization method for set data.”
paper / peterchapman / Feb 20
Commercial driving risk assessment is often biased due to a lack of contextual awareness. This paper proposes Expert-centered Driver Assessment (EDA), a methodology integrating expert input on contextual factors (e.g., weather, traffic) into AI-driven driver assessment systems. EDA uses fuzzy sets to model expert uncertainty and variance in factor influence, leading to fairer and more accurate risk evaluations.
commercial-drivingdriver-assessmentmachine-learning-fairnessai-moderationfuzzy-logichuman-computer-interaction
“Traditional intelligent commercial driver-assessment systems produce biased and inaccurate assessments due to their inability to consider contextual factors.”