Support and Partnerships
Project presentation for partners and institutions
A technological solution for protecting victims of digital violence
Context: The Failure of Current Responses
A Systemic Crisis
Non-consensual dissemination is a form of mass violence, amplified by technology. Victims face a cycle of powerlessness:
- Virality and infinite duplication of content.
- Unsustainable mental and procedural burden.
- Lack of active monitoring and preventive solutions.
- Insufficient technical and psychological support.
Key Figures
+800% increase in pornographic deepfakes (2018-2023).
1 in 10 young women in Europe is a victim of non-consensual dissemination.
98% of deepfake victims are women.
Our Value Proposition
ShieldHer is not just a reporting tool. It is a proactive and automated protection ecosystem, designed to shift the balance of power.
- PROACTIVITY: We detect content without prior reporting, often before the victim is even aware.
- AUTOMATION: Our technology processes millions of pieces of content per day, a scale impossible to achieve manually.
- CONFIDENTIALITY: Our "Privacy by Design" architecture ensures absolute security and anonymity.
- EMPOWERMENT: We give victims back control over their image and the means to act effectively.
Modular Technical Architecture
A scalable, resilient infrastructure designed for performance.
Crawlers (CPU Units)
Massive and specialized web crawling to discover and monitor source sites.
Processing (GPU Units)
Media analysis: NSFW classification, facial detection, pHash calculation.
Secure Embeddings
Transformation of faces into encrypted, non-reversible, and anonymous fingerprints.
Databases
Indexing of URLs, fingerprints, and metadata for fast searching.
Each module is independent and can be scaled separately, ensuring the system's smooth evolution.
Our Artificial Intelligence Models
We develop and refine our own models for maximum precision and relevance, ensuring our technological sovereignty.
Classification model to identify potentially intimate content with high sensitivity.
Algorithms optimized for robustness (angles, low resolution) and the creation of secure embeddings.
Detection of duplicates and similar content (cropping, filters) to combat replication.
Development of solutions to identify manipulations and synthetic content.
At the Core of the Project: Confidentiality & Security
Our "Privacy by Design" architecture is non-negotiable. It is based on strict principles to guarantee the absolute protection of victims.
- Zero human access: No operator can view the content or results. The system is 100% autonomous.
- No image storage: Only non-reversible digital fingerprints (embeddings) are stored. A face cannot be reconstructed.
- End-to-end encryption: Data is encrypted in transit and at rest, with industrial-grade security protocols (TPM, etc.).
- Exclusive victim control: Only the concerned person, after a secure biometric verification (embedded KYC), can access their results.
Use Cases and Beneficiaries
Who do we help?
- Victims of revenge porn and content leaks.
- Targets of sextortion and blackmail.
- People whose image is used in deepfakes.
- Victims of assault whose images are disseminated.
Who are our partners?
- Victim support associations, who can rely on our tool.
- Platforms and hosting providers, to help them comply with the DSA and protect their users.
- Institutions and competent authorities, as a technological detection brick.
Advantages for our Partners
Collaborating with ShieldHer means committing to concrete action and benefiting from unique expertise.
- For associations: A powerful tool to relieve victims of the technical burden of searching, allowing you to focus on psychological and legal support.
- For platforms: A turnkey solution to comply with regulations (DSA), proactively moderate illegal content, and enhance your ethical image.
- For institutions: A cutting-edge technological brick for monitoring and analysis, which can help characterize dissemination networks without handling sensitive data.
The Collaborative Support Pathway
We offer a seamless pathway where technology and human support are complementary.
- Referral: A victim contacts you. You inform them about ShieldHer and direct them to our secure platform.
- Onboarding: The victim independently performs a biometric identity verification via our mobile app. This process is designed to be simple and reassuring.
- Detection: The system compares their facial fingerprint to our database and presents any matches in a personal and secure space.
- Action & Support: The victim can initiate takedown requests. Your role in providing psychological and legal support then becomes paramount, freed from the burden of technical research.
Key Performance Indicators (KPIs)
Technical & Performance
- Content analyzed/day (>30M)
- Average detection time (<24h)
- False positive/negative rate
Human Impact
- Victims supported
- Takedown requests launched
- Average takedown time
Coverage & Dissemination
- NSFW sites indexed
- Partner platforms
- Facial clusters detected
Strategic Roadmap
Phase 1 (0-9 months)
Initial infrastructure deployment, consolidation of AI models, and establishment of first test partnerships with volunteer associations.
Phase 2 (9-18 months)
Infrastructure expansion (x10 scalability), launch of the secure mobile application, and signing of agreements with major platforms.
Phase 3 (18-36 months)
Start of R&D on non-public web access (autonomous AI agents), scientific publication, and positioning as a technological standard.
Governance and Ethical Framework
Trust is our most valuable asset. Our structure is designed to preserve it.
- Non-profit Status (French 1901 law): Guarantee of our non-profit mission and transparent governance.
- Controlled Intellectual Property: The core technology is protected to prevent any commercial drift or appropriation.
- Advisory Board: A committee of experts and association representatives will be established to ensure ethical compliance.
- Transparency: Regular activity reports will be published for our partners and funders.
Business Model & Sustainability
Our hybrid model aims for financial autonomy to guarantee a free service for victims in the long term.
Funding Sources
- Public & institutional grants
- Donations & corporate sponsorship
- Solidarity-based B2B model with platforms
Parallel Commercial Structure
A separate entity will offer paid services to professionals (copyright protection), and will commit to donating at least 50% of its profits to the association to ensure its sustainability.
Needs & Next Steps
Initial funding is crucial to deploy the infrastructure and demonstrate our impact on a large scale.
Key Material Needs
- Compute servers (CPU/GPU)
- Storage infrastructure (NAS)
- Network and security
Estimated initial budget: ~€80k for a reactive version (x10).
Strategic Human Needs
- Administrative coordination
- Fundraising
- Partnership development
Let's discuss our future collaboration
Contact us for a demonstration or to discuss our partnership terms.
contact@shieldher.org