Video summary

How to Gather Information on Anyone with OSINT

Main summary

Key takeaways

Educational

Main ideas / concepts conveyed

  • Ubiquitous surveillance + oversharing

    • Modern surveillance is portrayed as widespread and increasingly commercial (the “surveillance economy”).
    • People voluntarily publish personal data in real time (e.g., Snapchat/Instagram stories), creating a “digital catalog.”
    • Anything done online can leave a trail, even if the person tries to be private.

  • What OSINT is

    • OSINT (Open-Source Intelligence) is defined as the practice of collecting, analyzing, and documenting publicly available information to answer a specific intelligence question.
    • OSINT can be used to answer questions such as:
      • identifying activity behind online content
      • connecting identities
      • locating information tied to online files/communications
    • The narrator emphasizes OSINT can be used for legitimate ends (journalism, security research, law enforcement) but is also abused (harassment, stalking, doxxing).

  • Why OSINT exists

    • Information is no longer scarce: data is abundant due to public posting and platform behavior.
    • Companies have incentives to track/store/archive data; much of it becomes public or accessible.

  • OSINT history and evolution (high-level timeline)

    • Intelligence gathering predates “OSINT” as a term, rooted in conflict and strategy.
    • Key “analog” leaps:
      • Mass printing / newspapers (information becomes easier to collect)
      • Radio (broad access to broadcasts; military monitoring emerges)
      • World War II: the U.S. establishes a Foreign Broadcast Monitoring Service to monitor/record/transcribe foreign broadcasts
      • Cold War: “information war” conditions; OSINT grows alongside other intelligence disciplines
    • The term becomes “official” around 1990 (citing CIA usage), then expands with the internet and social media.

  • The narrator’s framing: OSINT as “puzzle pieces”

    • OSINT is portrayed as assembling a puzzle from many small data points.
    • Even “minor” behaviors (posting times, phrasing, slang) can contribute to identity resolution.

Methods / instruction-like guidance presented

1) Location finding from photos (geolocation / “visible OSINT”)

  • Use visual cues to narrow down location

    • Identify the language of signs (e.g., English) and distinguish between country-specific sign styles/colors.
    • Use architecture and likely city comparisons to shortlist possible cities.
    • Look for flags/logos on-site if visible.
    • Use other street-level signs (e.g., “no parking”) as additional constraints.

  • Use mapping and matching

    • Search major storefronts/landmarks (example approach: check every store chain location citywide, then compare backgrounds).
    • Use Street View to find the exact matching building/background.

  • Use landmarks for faster matching

    • Identify distinctive statues/landmarks in the image.
    • Compare adjacent background buildings to match the exact camera position.

  • Warning: even private or natural settings can be geolocated

    • Trees and shrubs can be identified to estimate location.
    • Example methodology:
      • identify tree species (example: Acer platanoides)
      • identify shrub species (example: Prunus serasus)
      • use QGIS plus a Python script to narrow candidates
      • manually check remaining candidates until the exact spot is found

  • Practical takeaway

    • Don’t assume privacy even “in the woods” or “inside your house.”
    • Encourage readers to test geolocation on their own public photos as an educational exercise.

2) Metadata extraction and stripping (“invisible OSINT” risk control)

  • Understand metadata types

    • File metadata: may include compiler paths, timestamps, environment variables, and potentially tokens/keys (depending on how the file was produced).
    • Image metadata (EXIF): may include exact GPS coordinates, date/time, device info, camera settings.
    • Video/file metadata: may include creation/editing history and other embedded details.

  • Key tools / tactics mentioned

    • Use ExifTool to extract metadata.
    • Use an exif/metadata stripping tool (the narrator mentions “exif tool” generally; later references ExifTool or similar) to remove metadata before uploading/sending.

  • Platform behavior cautions

    • Some platforms strip metadata inconsistently or only sometimes:
      • The narrator claims some social media may strip location data, but may still harvest metadata internally (advertising, subpoenas, etc.).
      • Reddit (desktop): claimed to sometimes retain exact location data.
      • Discord: claimed that uploading as a file (not via image upload button) may preserve metadata; historically video stripping was inconsistent.
      • Telegram: described as having similar issues to Discord.
      • LinkedIn private chats: claimed to not strip image metadata.

  • Threat framing

    • Metadata can be enough to determine coordinates and identify origin (who/where content came from).
    • Examples used to illustrate impact:
      • “Burger King foot lettuce” case: metadata not stripped on upload; later discovered via coordinate/time extraction.
      • “Epstein cell video” claim: edited status inferred through metadata/tools used to edit/export and stitched sources.
      • “I Know Where Your Cat Lives”: public geotagged cat images mapped using accessible location data.

3) Recovering deleted or altered content (internet archiving)

  • Use the Wayback Machine

    • Purpose: access historical snapshots of websites.
    • Covers:
      • pages that no longer exist
      • modified content
      • video/audio archives
      • older software/games/application snapshots
    • Practical use case: if someone deletes posts, archived snapshots may still preserve them.
    • The narrator warns (with a joke) that archived content can resurface long after posting.

  • Alternatives

    • Other archiving services are mentioned as comparable options.

4) Identity resolution from behavioral “patterns”

  • Collect pattern data

    • Posting/online times → suggests timezone and routine.
    • Reused usernames/images/phrases across accounts.
    • Dialect and regional language patterns (examples include “pop/coke/minerals/fizzy drink”).
    • Writing/typing habits:
      • consistent misspellings
      • recurring phrasing/idioms
      • recurring ideological markers across accounts

  • Recognize stylometry and voice features

    • Even with voice changers, inflection/word emphasis/tone/accent can persist as patterns.

  • Example case used

    • Ted Kaczynski (Unabomber): linked to manifesto writing style/unique phrasing recognized by a relative (including an unusual phrasing of a common proverb).

  • Core conclusion

    • Deception doesn’t fully erase patterns—changing patterns too little leaves identifiable footprints.

5) Documentation / evidence standards (for professional/legal usefulness)

  • OSINT must be proveable and auditable

    • Findings should be documented so others can verify and trust them.

  • Documentation workflow described

    • Form an intelligence question
    • Create a hypothesis
    • Collect and document evidence that corroborates the hypothesis
    • Use sources like Wayback Machine snapshots plus independent corroboration
    • Build a timeline and summary of findings

  • Example (fictional scenario)

    • Investigate a hypothetical company’s “no logs” claim before/after an incident using archived versions and other news sources.

6) Social media OSINT workflow + tools (identity resolution focus)

  • Primary task

    • Identity resolution: determining whether two accounts belong to the same operator.

  • Techniques

    • Pattern recognition: activity times, image reuse, unique phrases, usernames.
    • Account linkage even via “smokescreen” accounts—many patterns are subconscious.

  • Tooling examples mentioned

    • Sherlock for username tracking across platforms.
    • Other tools mentioned for:
      • Facebook account lookup/recovery
      • Instagram data collection
      • Reddit behavior/keyword extraction
      • Snapchat map scraping
      • Telegram monitoring tools (described as mass-surveillance style)

7) Dark web OSINT (high-level only; avoids illegal instructions)

  • Purpose framing

    • Dark web OSINT is described as mostly for:
      • security research
      • cyber attack awareness
    • It’s emphasized as difficult and rarely used to directly identify users due to onion-network anonymity.

  • Process concept

    • Scrape dark web sources → validate information heavily (example: validation might be treated as “90%” of the effort).

  • Important boundary

    • The narrator refuses to provide explicit instructions for accessing illegal marketplaces or harvesting stolen data.

8) Google dorking (powerful search operators)

  • General claim

    • Search operators can massively elevate OSINT queries.

  • No deep how-to

    • The narrator points to a separate video for details.

9) Defense / risk reduction: data minimization and OPSEC

  • Data minimization

    • Keep social media presence minimal.
    • Don’t overshare photos “for approval.”

  • OPSEC

    • The narrator advises improving operational security (explicitly avoiding deep detail, referencing an OPSEC video).

  • Platforms’ ownership claim

    • Once uploaded to platforms, photos belong to the platform per terms of service, enabling repurposing (including deepfakes)—presented as a privacy/consent concern.

  • Ethical guardrails

    • Use OSINT responsibly; illegality and unethical outcomes depend on how data is used.

10) Future outlook: automation + validation

  • AI impact

    • Prediction of an “arms race”:
      • automation will make doxing more capable
      • defenders will also be able to generate fake evidence quickly
    • The “hardest part” shifts from collecting to validating information.

  • Surveillance scaling risk

    • Fear that AI will drive further mass surveillance, increasing concern about the tracking trajectory.

Speakers / sources featured (as requested)

  • Main speaker (narrator): The host, referred to as “Zuma” (also “Zuma on YouTube”).
  • Guest/source referenced in the content:
    • Sun Tzu — The Art of War (quoted)
    • CIA officer / CIA — referenced as the origin of the term “OSINT” around 1990
    • U.S. Foreign Broadcast Monitoring Service — referenced as a WWII-era monitoring effort
    • Ted Kaczynski (“Unabomber”) — referenced via manifesto style evidence
    • David Kaczynski — referenced as the person who recognized writing style
    • Burger King foot lettuce case (“Burger King” employee) — referenced as an example
    • U.S. Department of Justice — referenced regarding the “Epstein cell video” claim
  • Notable third-party programs/tools named:
    • Wayback Machine
    • ExifTool
    • QGIS
    • Python
    • Sherlock (username tracking)
    • Tor / Tor Browser (for onion sites)
    • Google dorking (Google search operators)

Original video