AI-assisted evidence infrastructure for atrocity accountability.
Stratos builds secure, human-led AI systems that help investigators and lawyers transform testimony, open-source material and multilingual digital evidence into protected records, legal dossiers and policy action.
The accountability problem is now a data problem as well as a legal oneAtrocity evidence is often dispersed, multilingual, time-sensitive and dangerous to handle. Testimony may sit with survivors and families. Digital traces may be scattered across platforms. Legal thresholds vary by country. Public advocacy can expose witnesses if evidence is not managed carefully.
The Stratos Evidence Lab helps close that gap. It gives legal teams a secure way to structure evidence, identify corroboration leads, map actor networks, protect sensitive sources and prepare material that prosecutors, sanctions units, legislators and international bodies can use.
What we doFrom fragmented evidence to accountable action
We use multilingual AI tools to support search, translation workflows, entity extraction, clustering, redaction flags and review queues.
We collect and manage testimony and submitted evidence through consent menus, access controls, risk assessments and chain-of-custody metadata.
We link people, institutions, incidents, dates, locations, legal theories and corroborating material so investigators can see patterns without losing source control.
AI-assisted triage.
Legal dossier assembly.
We help lawyers prepare structured evidence tables, chronologies, sanctions packages, prosecutor referrals and policy annexes.
Governance and audit.
We operate under written rules for human review, safeguarding, data minimisation, model use, legal sign-off and independent technical review.
Responsible AI
AI supports the evidence process. It does not decide the case.
Stratos uses AI to assist with scale, structure and consistency. AI can help search large collections, suggest tags, find repeated entities, identify missing information and flag details that may expose a witness. It does not determine guilt, decide legal liability, rank survivor credibility or approve public naming.
Every legal conclusion, public allegation, witness-risk decision and filing strategy remains subject to human verification, safeguarding review and counsel approval.
· Human review for all outputs used externally.
· Child and survivor protection before publicity.
· Role-based access, encryption and audit logs.
· Donor firewall around legal case selection.
· Independent review of AI governance and data protection.
DonorsFund public-interest technology for legal accountability
A grant to the Stratos Evidence Lab will fund a distinct AI and data workstream serving two live accountability campaigns. Donors can support the 100-day prototype, a 12-month cross-project AI and evidence workstream, or a project-specific deployment in Ukraine or Iran.
The first phase will build the secure architecture, evidence schema, AI-assisted review workflows, pilot outputs and governance documents needed to scale the Lab responsibly.
· Fund the 100-day Evidence Lab sprint.
· Support the 12-month cross-project AI and evidence workstream.
· Back the Ukraine child-transfer evidence graph.
· Back the Iran testimony analysis and legal triage system.
Technical Partners
Stratos is seeking technical partners with experience in secure data architecture, multilingual NLP, knowledge graphs, digital evidence, model governance, information security and human-rights documentation. The partnership model is designed to combine technical capability with legal accountability expertise and independent governance review.
Partner capabilities
· Secure evidence architecture and data protection.
· Knowledge graph and entity-resolution systems.
· Multilingual NLP, semantic search and review queues.
· Human-in-the-loop AI evaluation and audit.
· Digital security, threat modelling and incident response.
The next generation of accountability work will require more than public reports.
It will require secure evidence systems, legal discipline, technical capability and governance strong enough to protect the people whose evidence makes justice possible. Stratos Evidence Lab is built for that task.