Project Overview
The Five Eyes AIR Dashboard is a public-facing research and education platform developed to support the study of Algorithmic Immersion Radicalisation (AIR) across Australia, Canada, New Zealand, the United Kingdom, and the United States.
The dashboard is based on the AIR framework developed by Kelly W. Sundberg and Alannah Reynaud. AIR examines how digital platforms, recommender systems, online communities, and repeated exposure to identity-reinforcing content may contribute to ideological escalation in contemporary security environments.
AIR is architecture-driven rather than ideology-specific. It does not argue that technology alone causes radicalisation, extremism, coercion, or violence. It treats online architecture as an accelerant that may shape the speed, depth, intensity, and social reinforcement of ideological immersion.
The dashboard is descriptive, comparative, and educational. It is not a diagnostic tool, a threat-assessment tool, or a predictive model.
At the centre of the AIR framework is a four-stage cycle: Curation, Immersion, Reinforcement, and Mobilisation. The cycle may arrest, recur, or intensify; mobilisation is a possible outcome, not an inevitable one.

The AIR project distinguishes lawful democratic engagement from coercive ideological violence and mobilisation (CIVM), and from terrorism or violent extremism. Lawful protest, advocacy, criticism, religious expression, political disagreement, and democratic participation must not be conflated with extremism.
CIVM is an analytical category used for research. It is not a legal classification, a proposed offence, or a substitute for criminal-law thresholds.

The AIR framework is situated in a longer history of algorithmic recommendation, mobile connectivity, social media, encrypted group communication, short-form video, livestreaming, and AI-enabled content environments. This timeline provides conceptual background; it is not a live case trend and does not claim that technological change alone causes radicalisation or violence.

Current case-pattern text loads from the active case dataset.
The Social Context Data Master provides national-level comparison indicators for the Five Eyes countries. These indicators help users understand broader environments in which AIR patterns are studied, including digital immersion, social media use, trust, public safety, loneliness, mental health, demographic change, and related social conditions.
Social-context data are not used to explain individual cases, diagnose societies, or claim that a country’s social-media use, crime rate, trust level, or demographic profile caused any AIR case.
The Five Eyes AIR Dashboard is not a static report. As case data are reviewed, new cases are added, social-context sources improve, and dashboard methods are refined, public-facing charts update from the underlying data files.
Methods and Limits
This dashboard supports descriptive and comparative analysis of AIR-related cases and national-level social-context indicators across the Five Eyes countries. It is not a diagnostic, threat-assessment, or predictive tool.
The Five Eyes AIR Dashboard is a public-facing research and education platform for exploring Algorithmic Immersion Radicalisation (AIR) case patterns and related national-level social context.
Cases are included when public sources support classification as a terroristic attack, thwarted terrorist plot, or coercive ideological violence and mobilisation (CIVM) incident within Australia, Canada, New Zealand, the United Kingdom, or the United States. The dataset is a living research instrument and may change as sources improve.
Terroristic Attack refers to an executed ideologically framed violent attack. Thwarted Terrorist Plot refers to a disrupted or prevented plot. CIVM refers to coercive ideologically framed mobilisation below, beside, or outside formal terrorism thresholds. CIVM is an analytical category, not a criminal offence, legal classification, or policy recommendation.
Higher, Moderate, Lower, Nil, and Under Review describe the apparent public-source-supported role of online or digitally mediated immersion. AIR level is not a diagnosis, risk score, or legal finding.
Key variables include country, year/date, case category, AIR level, ideology / motivation, target group, target location, weapon / method, public summary, immigration / citizenship reporting category, offender religion, and sex / gender.
Social-context indicators provide national-level background for comparison. They use public terminology such as Context Area, Topic, Indicator / Measure, Raw Value, Five Eyes Average, and Standardized Five Eyes Comparison. They are not causal explanations for individual cases.
Missing values are not treated as zero. Unavailable values remain unavailable. The 25 and Older age category is shown only where populated; All Ages is not silently substituted for 25 and Older.
Five Eyes Average is shown only where valid. Standardized comparison may use z-score values supplied by the Social Context Data Master for visualization, but this is not a risk score. Raw values remain available in public downloads.
Test Yourself compares user-entered time estimates with available benchmark values and identifies platform exposure features. It does not diagnose addiction, radicalisation, risk, or behaviour.
Records use source_id values to connect dashboard data to visible references where possible. Some sources are retained for traceability without clickable links when no usable public URL is available or access is restricted.
Limitations include public-source constraints, evolving case information, differences across national data systems, social-context comparability limits, no causal inference, no individual behaviour prediction, and no clinical or threat-assessment function.
Public Downloads
These curated downloads are public-facing extracts. Internal governance fields, audit-only fields, security files, and admin-only metadata are not included in ordinary public downloads.
Raw Value is retained in public social-context downloads. Full technical runtimes remain in the site data folder for dashboard operation but are not presented here as ordinary public downloads.
We welcome evidence-based suggestions, corrections, and source recommendations.
References are grouped for readability. Sources without public links are displayed as plain text for traceability.