Home / OSINT Sources & Methods
All OSINT — No Classified Material

Intelligence Sources & Methods

Transparency is the foundation of credible open-source intelligence. This page documents every category of source we use, how we verify claims, how we assign confidence, and where our analysis has limits. No intelligence relationship. No leaked documents. No anonymous insiders. Just rigorous public source analysis.

Source Categories

Our 12 Source Categories

Every claim on this platform traces back to at least one of these twelve categories of publicly available information. We do not use classified, proprietary, or illegally obtained material under any circumstances.

01
DOC
Government Documents
Official publications from defense ministries, legislative bodies, and executive branches. The most authoritative source for confirmed programs, budgets, and stated policy.
DoD Budget Requests GAO Reports Congressional Testimony FOIA Releases NDAA Text DIA Threat Assessments
02
PAT
Patent Filings
Defense company patent applications reveal technical capabilities months or years before public demonstrations. Patent language is legally binding — companies do not patent what they haven't built or don't intend to build.
USPTO EPO (European) WIPO International CNIPA (China) Rospatent (Russia)
03
ACR
Academic Research
Peer-reviewed research from defense-affiliated universities and government labs provides technical ground truth. Academic output from Chinese, Russian, and Iranian defense institutions is particularly valuable for capability assessment.
arXiv IEEE Xplore Defense Technical Reports RAND Research DTIC NUDT Publications
04
SAT
Satellite Imagery Analysis
Commercial satellite constellation providers now offer sub-meter resolution imagery. Change detection analysis of defense facilities, test ranges, and production lines provides independent verification of development activity.
Maxar Technologies Planet Labs Airbus Defence BlackSky Sentinel-2 (ESA)
05
SEC
Defense Industry Filings
Publicly traded defense contractors are legally required to disclose material contracts, revenue segments, and risk factors. SEC filings, earnings calls, and investor presentations contain substantial intelligence on program status and spending.
SEC 10-K / 10-Q Earnings Call Transcripts Investor Presentations USASpending.gov SAM.gov Awards
06
ARM
Arms Trade Databases
Multilateral arms transfer databases track the movement of weapons systems between countries. These databases provide empirical grounding for proliferation analysis and are maintained by respected independent research institutions.
SIPRI Arms Transfers DB UN Register of Conventional Arms National Export Reports ACLED Janes Weapons DB
07
SOC
Social Media / OSINT
Conflict documentation on Telegram, X, and other platforms provides real-time ground truth on weapons deployment. Verified OSINT accounts, military unit channels, and geo-confirmed imagery are evaluated for authenticity before citation.
Telegram (Ukraine/Russia) X Defense Accounts Bellingcat Network GeoConfirmed OSINTdefender LinkedIn (Defense)
08
TTK
Think Tank Reports
Established security and defense policy research institutions produce extensive analysis drawing on expert networks, retired military professionals, and direct government relationships. We treat these as high-confidence secondary sources.
RAND Corporation CSIS Brookings Institution IISS Carnegie Endowment Hudson Institute
09
MED
Specialist Media
Defense trade publications employ journalists with direct access to programs, officials, and procurement processes. These outlets are essential for current-events tracking and serve as primary sources for contract announcements and program developments.
Breaking Defense Defense One Janes Defence Weekly The War Zone Defense News Aviation Week
10
INT
International Organizations
Multilateral bodies produce treaty texts, meeting records, working papers, and formal submissions from member states. These documents are essential for policy and legal analysis, and for tracking the evolution of international norms around autonomous weapons.
UN CCW Documents ICRC Reports NATO Public Docs EU Council Documents IAEA
11
TRD
Trade Shows & Conferences
Defense exhibitions are where new systems are publicly unveiled. Brochures, demonstration videos, and official presentations from major defense expos constitute primary source documentation of system specifications and program status.
AUSA Annual Meeting DSEI (London) IDEX (Abu Dhabi) Paris Air Show Zhuhai Airshow Eurosatory
12
STD
Technical Standards & Controls
Export control frameworks and multilateral technology control regimes document which technologies are considered militarily sensitive and track national export authorization patterns. These inform proliferation risk analysis.
Wassenaar Arrangement MTCR ITAR / EAR (US) EU Dual-Use Regulation NSG
Analytical Process

Our Methodology

A source is only as good as the process applied to it. These are the six principles that govern how we collect, verify, score, and publish analysis on this platform.

Rule 01
Multi-Source Verification
No factual claim is published on the basis of a single source. A minimum of two independent sources are required to support any specific claim — where "independent" means different organizations, different document types, or different collection methods. Where only one source exists, the claim is explicitly labeled as single-source and confidence is reduced accordingly.
Rule 02
Confidence Scoring System
Every prediction and major analytical claim carries an explicit confidence percentage. High confidence (75%+) requires strong multi-source evidence with documented trend lines. Medium confidence (50-74%) indicates plausible but uncertain trajectories. Low confidence (below 50%) signals predictions that are possible but dependent on contingencies. Scores are reviewed when new evidence emerges.
Rule 03
Source Hierarchy
Not all sources carry equal weight. Primary sources (government documents, patent filings, official statements) outweigh secondary sources (media reporting, think tank analysis). Where primary and secondary sources conflict, we note the discrepancy rather than resolve it arbitrarily. Social media OSINT is treated as indicative until corroborated by more authoritative sources.
Rule 04
Bias Acknowledgment
Our source base has structural limitations. We rely heavily on English-language sources, which creates blind spots for Russian, Chinese, Farsi, and Arabic-language primary material. We have no access to classified assessments, which means our analysis reflects what states choose to reveal — not necessarily their actual capabilities. We acknowledge these gaps rather than treat our analysis as comprehensive.
Rule 05
Update Frequency & Corrections
Major analytical pages are reviewed at minimum monthly. Predictions are updated promptly when resolving evidence emerges. Corrections are published in-line with the original content — we do not delete or silently revise. When a prediction is confirmed or failed, a timestamped evidence note is added documenting the resolution. We maintain a public correction log.
Rule 06
Analytical Boundaries
We do not speculate beyond what the available evidence supports. When analysis requires extrapolation, we label it explicitly. We do not publish targeting information, specific vulnerability assessments of critical infrastructure, or any content that could provide material assistance to an actor seeking to cause harm. Our audience is the informed general public and policy researchers — not operational planners.
Confidence Framework

How We Score Confidence

Confidence percentages on this platform are not arbitrary. They map to specific evidentiary thresholds. Here is exactly what each tier means.

75%+
High Confidence
Multiple primary sources. Documented trend line with consistent direction over 12+ months. Technical feasibility confirmed by published research. No significant countervailing evidence. The prediction is likely unless an unusual disruptive event occurs.
50-74%
Medium Confidence
At least two independent sources. Plausible trajectory with some countervailing factors. Outcome is dependent on 1-2 contingencies resolving favorably. Technical feasibility established but timeline uncertain. Considered the probable outcome, not the certain one.
Below 50%
Low Confidence
Limited or conflicting evidence. Outcome is possible but requires multiple contingencies. Technical feasibility is uncertain or unproven. We publish low-confidence predictions because falsifiable forecasting requires stating bold claims — we accept we may be wrong.

"Artificial Weapons is a public open-source intelligence platform. No classified, proprietary, or illegally obtained information is used in any analysis published on this site. All intelligence is derived exclusively from publicly available sources including government documents, academic publications, patent filings, commercial satellite imagery, defense industry disclosures, arms trade databases, and verified open-source reporting."

Artificial Weapons publishes for educational, research, and public interest purposes. Nothing on this platform constitutes military advice, investment advice, or legal guidance. Analysis reflects the views of the editorial team based on publicly available information and may be incorrect. Users are encouraged to consult primary sources independently. If you believe any content on this platform is inaccurate or improperly sourced, contact us directly.