Safety & Privacy by Design
- Does the tool respect user data privacy?
- Is there any risk that this tool could be weaponised by abusers if misused?
- Are there safety fallbacks, such as a quick-exit button?
#16DAYS of Activism against Gender-Based Violence
Code for Dignity is a 16-day hybrid hackathon running alongside the daily #16DAYS conversations. It turns community dialogue into open-source digital infrastructure built to defend women and girls online.
Ensure Justice · Protect Survivors · End Impunity
The problem
Tech-Facilitated Gender-Based Violence (TFGBV), including targeted doxxing, deepfake generation, and bot-driven hate campaigns, is scaling faster than standard content moderation can handle. Victims are routinely gaslit by tech interfaces, while corporate frameworks fail to provide actionable, localised protection or legally sound evidence mechanisms.
The campaign
#16DAYS OF ACTIVISM against Gender-Based Violence is an annual international campaign. It kicks off on 25 November, the International Day for the Elimination of Violence against Women & Girls, and runs until 10 December, Human Rights Day.
Lianna Adams has hosted it for the past four years to create awareness and hold solution-finding conversations that challenge violence against women and girls.
Code for Dignity runs in sync with those conversations. Each daily conversation is a live user-research module. Its last 15 minutes deliver a Dignity prompt that hackathon participants use to iterate, prototype, and build.
What to build
Build anything small that reduces digital harm to women, girls or children, or makes a tool safer by design. Improving a build you already have counts.
How it works
Daily conversations covering the four tracks of the #16DAYS OF ACTIVISM campaign.
Each conversation ends with a 15-minute Dignity prompt on the day's topic.
One small build in any no-code or low-code tool.
The build, a 3-minute demo, and a one-page safety note.
Builds are selected to be presented on December 10.
Live demos and showcase.
A public gallery, a briefing pack to platforms, and an archive in the Code for Dignity Commons.
The 16 days
12:00 PM ET on X Spaces + Discord . Open a day for its conversation angle, Dignity prompt, 10-minute task and solution-building prompt. Join the Discord server to take part.
Digital abuse is not a communications problem; it is a rights, safety, and equality problem. Explore how harassment, stalking, image abuse, doxxing, and gendered disinformation converge into offline risk.
Dignity prompt“What would a digital world look like if a woman’s safety did not depend on her willingness to disappear from it?”
Map one survivor journey across platforms, devices, services, and offline consequences; mark every point where harm can escalate.
Build a TFGBV harm map: a shared, privacy-preserving taxonomy that helps teams identify abuse patterns, escalation risks, and affected groups without collecting identifiable survivor data.
Move beyond “use strong passwords.” Examine metadata, shared calendars, tagged photos, delivery records, fitness apps, public registries, and cross-platform inference.
Dignity prompt“What is the minimum amount of my life that an app truly needs to know?”
Audit a common app or service: list data collected, inferred, retained, shared, and visible to another account holder.
Prototype a footprint minimiser: a checklist or browser extension that flags high-risk exposures, suggests privacy actions, and never requires uploading personal data.
Safety tools often recreate coercive control by monitoring survivors “for their own good.” Ask: who holds the power, who can see the data, and who can turn the tool off?
Dignity prompt“How do we protect someone without making them feel watched, managed, or disbelieved?”
Redesign a hypothetical “safety app” around consent: what must be optional, encrypted, local-only, time-limited, or survivor-controlled?
Build a consent-first safety profile: local, encrypted templates for trusted contacts, check-in preferences, emergency escalation, and automatic data expiry.
A safe digital space is not only a website; it must work for people with shared devices, limited data, fear of discovery, disability, or unstable connectivity.
Dignity prompt“What does safety look like when privacy cannot be assumed at home?”
Sketch three user contexts: shared phone, no personal email, low bandwidth. Identify failure points in a typical support-service journey.
Prototype a low-friction digital safe-house entry point: pseudonymous access, quick exit, minimal data, offline-capable resources, and clear “what happens to what I type” notices.
Abuse is often narrative warfare: reputational attacks, false accusations, sexualized rumors, and coordinated pile-ons designed to silence women in public life.
Dignity prompt“How do we counter a lie without amplifying it or putting the targeted person on trial?”
Analyze a fictional coordinated-harassment scenario. Identify signals of coordination versus ordinary disagreement, and the risks of automated labeling.
Build a narrative-risk toolkit: detection of coordinated patterns, draft response templates, evidence-preservation guidance, and support pathways, without publishing accusations or identifying targets.
Synthetic media turns humiliation into scalable abuse. The issue is not only detection; it is rapid takedown, evidence preservation, victim support, and preventing re-victimization.
Dignity prompt“If a fake image of me appeared tomorrow, what would I need in the first hour, day, and week?”
Create a first-72-hours response map: document, report, request removal, seek support, and protect mental health, without requiring the survivor to repeatedly explain the abuse.
Prototype a synthetic-abuse response companion: consent-based evidence capture, platform-report preparation, takedown tracking, and links to legal/psychosocial support.
Most products are designed for a calm, well-resourced user. Ask what happens when a user is panicked, stalked, coerced, or using a device controlled by someone else.
Dignity prompt“What does this product do when the person using it is afraid?”
Threat-model one feature: who could misuse it, who could be harmed, what data could be exposed, and what default would reduce that risk?
Produce a safety-by-design checklist and apply it to a prototype: risk assessment, privacy defaults, abuse scenarios, human escalation, and transparency.
Survivors are often asked to become forensic investigators. Explore how evidence can be preserved credibly while reducing repeated disclosure, disbelief, and institutional harm.
Dignity prompt“What would justice feel like if I did not have to prove my humanity before my case?”
Map the evidence journey from disclosure to court. Identify where evidence is lost, where the survivor must repeat themselves, and where bias may enter.
Build a digital-evidence preservation prototype: timestamped, hash-verified, exportable evidence bundles with survivor-controlled sharing and clear legal-context guidance.
Moderation is not just deletion. Failures include under-enforcement, over-removal, opaque appeals, language bias, and inconsistent treatment across regions and identities.
Dignity prompt“What would fair moderation look like if the person reporting had the same standing as the person reported?”
Compare three moderation outcomes for the same abusive content: removal, no action, and account restriction. Identify transparency and appeal gaps.
Prototype a moderation accountability layer: structured reports, plain-language decisions, appeal tracking, and anonymised pattern reporting for civil-society oversight.
Moderators, helpline workers, and activists absorb trauma. Burnout is not a personal failing; it is a design and staffing failure.
Dignity prompt“Who cares for the people who hold other people’s pain?”
Design a shift protocol for a responder: intake limits, debrief prompts, escalation rules, and boundaries for after-hours contact.
Build a responder wellbeing module: caseload signals, debrief prompts, peer-support routing, and configurable boundaries, without exposing sensitive case content.
Prevention must not put responsibility on girls to avoid abuse. Focus on age-appropriate agency, trusted adults, platform accountability, and education that does not shame victims.
Dignity prompt“How do we teach girls to navigate digital spaces without teaching them that abuse is their fault?”
Rewrite a victim-blaming safety message into an agency-based, rights-based message for a specific age group.
Prototype an age-appropriate digital resilience toolkit: scenario-based learning, trusted-adult connection, reporting support, and privacy controls designed with young people.
Bystanders often want to help but fear retaliation or do not know what helps. Explore bystander power, collective documentation, and support that does not endanger the target.
Dignity prompt“What can a community do in ten minutes that makes a targeted person less alone?”
Design a bystander decision tree: witness, assess risk, support privately, document safely, escalate, or disengage.
Build a community response toolkit: private support messages, safe documentation prompts, resource routing, and coordination features that avoid public pile-ons.
A hackathon can accidentally reward flashy demos. Ask how to measure reduced harm, increased confidence, improved reporting, better institutional response, and survivor trust.
Dignity prompt“What evidence would tell us this tool made someone safer, not merely that people used it?”
Define one outcome, one indicator, one data source, and one ethical guardrail for a prototype.
Create an impact measurement canvas: safety outcomes, survivor experience, accessibility, privacy, unintended harms, and community accountability.
AI trained on the internet inherits gendered stereotypes, victim-blaming narratives, and unequal language coverage. Ask who is represented, who is misclassified, and who bears the error.
Dignity prompt“If this system makes a mistake about me, who will notice, who will correct it, and who will be harmed?”
Audit a proposed AI feature: inputs, training assumptions, error cases, human oversight, appeal path, and data-minimization requirements.
Build an AI harm and bias review card: intended use, excluded uses, known biases, human-in-the-loop requirements, data retention limits, and redress process.
Submissions are due today: the build, a 3-minute demo, and a one-page safety note.
Technology cannot substitute for law, enforcement, platform accountability, and survivor-led policy. Ask how code can create evidence, visibility, and pressure for institutional change.
Dignity prompt“What must change outside the app for justice inside the app to be possible?”
Translate one prototype into a policy ask: what rule, standard, disclosure, or duty would make the intervention systemic rather than heroic?
Produce a prototype-to-policy brief: problem, evidence gap, recommended safeguard, responsible institution, implementation steps, and survivor-informed success measures.
Human Rights Day. All builders present their build and receive live feedback and questions from the audience.
Problem, users, safety assumptions, data practices, limitations, maintenance plan, and community governance.
A public gallery, a briefing pack to platforms, and an archive in the Code for Dignity Commons.
December 10 · Live Demos Showcase
Submit by December 8: the build, a 3-minute demo, and a one-page safety note. Selected builders present live on December 10.
Design principles for builders
No scraping, public shaming, location exposure, facial recognition, or retention of sensitive evidence without explicit, revocable consent.
Survivors decide what is stored, shared, escalated, or deleted.
Assess risks before features; make privacy and safety defaults, not optional add-ons.
Design for girls, disabled women, LGBTQ+ people, migrants, rural users, low-bandwidth users, and people facing multiple forms of discrimination.
AI can triage or explain; it should not make final decisions about a person’s safety, credibility, or case.
Open-source the non-sensitive infrastructure, while keeping survivor data, threat intelligence, and case details closed.
Mentors
Strategy Director, Firsthand
Founder & CEO, ikwe.ai
Founder & CEO, DigiArmorX
Social Protection Specialist
Founder, LeeLoo AI and École H (AI school for neurodivergent kids). Works at the intersection of public & international law (FR-US).
Questions
Join in
Come to listen, come to build, or both. Conversations run daily at 12:00 PM ET on X Spaces + Discord, November 25 to December 10.