Research map / Large language models / AI and society
Making Transparency Advocates: An Educational Approach Towards Better Algorithmic Transparency in Practice
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Ran an open-source educational workshop on algorithmic transparency with 27 professionals in news/media and tech startups, measuring literacy and advocacy willingness via surveys and interviews.
Problem
Research progress in explainable AI has not translated into organizational implementation of algorithmic transparency; organizations often undervalue or are unaware of transparency methods.
Approach
Create transparency advocates (motivated, "tempered radical" employees) through a free, domain-customized educational workshop, and characterize advocacy as conversational, implementational, and influential, with willingness varying by professional domain.
Key results
- Post-workshop survey means rose across all constructs, e.g., General Understanding of Algorithmic Transparency from 5.00 to 7.79 on a 10-point scale (n=15).
- Willingness for advocacy in professional life rose from 4.14 to 4.71 on a 5-point scale.
- Four participants (in the interviews, P2, P3, P4 and a survey respondent) reported advocacy actions in the days following; one spoke up at an organization-wide AI strategy meeting.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Attend the algorithmic transparency workshop | Workshop participants · Algorithmic transparency workshop participants (n=15) | general understanding (10-point mean): 5.00 -> 7.79 | Helped |
| Attend the algorithmic transparency workshop | Workshop participants · Algorithmic transparency workshop participants (n=15) | benefits and purpose (3-point mean): 2.00 -> 2.64 | Helped |
| Attend the algorithmic transparency workshop | Workshop participants · Algorithmic transparency workshop participants (n=15) | stakeholders (3-point mean): 1.64 -> 2.57 | Helped |
| Attend the algorithmic transparency workshop | Workshop participants · Algorithmic transparency workshop participants (n=15) | tensions between goals (3-point mean): 1.71 -> 2.50 | Helped |
| Attend the algorithmic transparency workshop | Workshop participants · Algorithmic transparency workshop participants (n=15) | willingness for advocacy: professional life (5-point mean): 4.14 -> 4.71 | Helped |
| Attend the algorithmic transparency workshop | Workshop participants · Algorithmic transparency workshop participants (n=15) | willingness for advocacy: personal life (5-point mean): 3.71 -> 4.43 | Helped |
Surprising findings
- Startup professionals often cannot afford to prioritize transparency despite having technical knowledge, while news and media professionals already value it but lack implementation tools.
- Each interviewee seemed to have a personal barometer for when transparency is necessary, with 'unwritten rules' guiding disclosure.
Limitations
- Small sample size (27 participants; 15 with paired surveys) limits generalizability and statistical analysis.
- Voluntary participation likely biased toward individuals already inclined to advocate.
- Participant drop-off in the online workshop.
Experimental setup
Datasets: pre- and post-workshop surveys (n=15) semi-structured interviews (7 participants)
Metrics: pre/post survey means and standard deviations on 6 construct thematic analysis themes from interviews
Builds on: tempered radicals stakeholder-first approach to algorithmic transparency model cards datasheets for datasets
Closest papers by meaning
- Thesis Summary: Operationalizing User-Inclusive Transparency in Artificial Intelligence Systems (AAAI 2024 · similarity 0.83)
- Advancing Research on Equitable AI Education Through a Focus on Implementation: Insights from a Middle School Computer Vision Module Beta-Test (AAAI 2025 · similarity 0.83)
- Learning About Algorithm Auditing in Five Steps: Scaffolding How High School Youth Can Systematically and Critically Evaluate Machine Learning Applications (AAAI 2025 · similarity 0.83)
- The Foundation Model Transparency Index (NeurIPS 2026 · similarity 0.83)
- Position: Bridge the Gaps between Machine Unlearning and AI Regulation (NeurIPS 2025 · similarity 0.81)
- We Are AI: Taking Control of Technology (AAAI 2025 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.