This Charter outlines SOLV’s core ethical standards for our work. SOLV is built on the principle that AI should amplify human judgment, not replace it, making ethical safeguards essential. This Charter supports our consistent, transparent, and responsible decision-making practices throughout all stages of SOLV. We're dedicated to responsible AI use, data protection, and transparent decision-making to build trust-based relationships with all stakeholders impacted by our work.
SOLV is committed to follow the core ethical principles enclosed in this Charter. It was first drafted in April 2025 by SOLV researchers and has since been updated to its current version in January 2026.
This Charter takes influence from the guidance and language used in several AI best-practices documents, including the; United Nations Educational, Scientific and Cultural Organization (UNESCO) Recommendations on the Ethics of Artificial Intelligence (adopted by 193 countries), and the Organization for Economic Cooperation and Development (OECD) AI Principles and recommendations.
We do not analyze or store people’s personal identity information (PII), such as name, date of birth, address, gender, security information, etc. The data that SOLV extracts is focused on the values and positions stakeholders bring to public interest discussions and relevant projects.
We only process the minimum data necessary to establish project context.
Client data remains the sole property of the client and is never used to train models, shared across projects, or repurposed beyond agreed-upon use.
All engagements are governed by a Data Processing Agreement (DPA) that defines strict usage boundaries and ensures compliance with GDPR-aligned practices.
SOLV uses Third-Party AI for the specific purpose of structured data extraction from unstructured materials (identifying stakeholders and values from reports, PDFs, meeting notes, etc.). Documents are temporarily processed by select contracted AI providers (OpenAI, Anthropic) solely for extracting data (names, organizations, statements, positions) and natural language processing.
All outputs are immediately stored within SOLV’s secure digital environment. Project information is not stored in Third-Party AI systems. DPAs with all AI providers contractually prohibit model training or secondary use of Client data, ensuring that all strategic intelligence remains confined to the Client’s secure SOLV workspace. This technical one-way data flow prevents any retention or reuse by external providers, and all subsequent analysis, modeling, and risk assessment occurs entirely within SOLV.
SOLV does not automate decisions. SOLV employs AI as an assistive tool, not a decision-maker. Every AI-assisted outcome is explicitly traceable to the underlying evidence with direct access to original source materials. AI is used to identify patterns and structure information, while Client teams retain responsibility for validation, refinement, and final interpretation at every stage. Where AI-generated ratings or assessments are presented, SOLV provides clear explanations of the methods and assumptions used to reach those conclusions. Clients maintain full visibility and access to what data was analyzed, how it was structured, and the reasoning underpinning all outputs.
SOLV is designed as a transparent, auditable system.
We remain accountable for SOLV and its possible weaknesses and are therefore focused on continuous improvement. We ensure transparent processes to explain data analysis and the use of AI to project stakeholders. We integrate stakeholder feedback loops into our partnerships to strengthen our work and enhance our impact. Section 7 outlines ongoing steps to advance these ethical guardrails (including creating an Independent Ethics Review board and releasing an annual Transparency Report).
We do not focus on averages of a community but highlight the active pluralism and value differences within groups. There is no singular opinion of a group, but a multi-faceted diversity of values and ideas. We apply these principles to all analysed data, ensuring data integrity and quality extraction. We avoid sampling bias by uploading all relevant data related to the project context. We do not manipulate data as we believe in mapping and identifying value clusters and coalitions as well as current and potential conflicts. We aim to make diversity visible and from there, enable solutions that address the diversity of viewpoints.
In alignment with our values of diversity and inclusivity, we are working on training our AI to note when bias is potentially evident in uploaded data. As all relevant documents to projects will be uploaded, we hope to teach our AI when to highlight potential biases. This lets us enhance our goal of SOLV supporting truly transparent decision-making processes.
We built this technology to enrich trust-based relationships between SOLV, Clients, stakeholders, and broader communities. This supports human decision making, at no point does SOLV technology make strategic decisions for you.
The SOLV platform scans documents, suggests patterns, and scores risks, but it is built so that the Client or stakeholder will curate the data, validate the pattern, and decide what to do with the information.
We require informed consent at all stages of live stakeholder engagement. We limit our public document analysis to materials and statements made explicitly available in the public domain, such as government consultations or published media commentary.
When doing a 360° listening tour or live engagements/facilitations, we aim and advise Clients to include as many voices as possible not only the ones aligned with those of the Client. We will not hide diversity within the data and we always advise Clients to listen as carefully as possible to as many voices and as diverse opinions as possible.
We advise our Clients to share findings from the collected data back to the stakeholders relevant to the project. Although the Client owns the data and decides how to use it, transparency with involved communities is key to building trust and ensuring a project's proper development.
We work with Clients who seek alignment, not manipulation. SOLV is especially designed to reduce conflict and enable pluralistic decision-making. We build our technology explicitly for situations addressing:
Democratic governance and cross-boundary coordination
Community-centered infrastructure planning
Transparent policy development
Trust-building in contested environments
We strive to identify and counter misuse of SOLV, as it is developed as a tool to facilitate long-term social impact and sustainable outcomes.
We reserve the right to decline projects that conflict with these values.
We continue to improve our practice to ensure quality data extraction and transparency. We are working on:
Independent Ethics Review by the end of 2026
External oversight for sensitive projects
ISO 27001 Certification by 2027
Information security management
Annual Transparency Reports, TBD
Public reporting on usage, compliance, and ethics
Additional monitoring mechanisms, TBD
Developing bias training for the platform, etc.