The Network / Evidence-checked reading trails

Start with one file.
Follow the evidence.

Each line between two files says why they connect, what the evidence is, and where the link stops.

7 trails · 29 investigations · 27 explained connections

Trail 01 / Surveillance

Who can track us?

Private cameras, commercial data and allied intelligence all lead to one question: who gets to search what has already been collected.

Case file / 01

Flock’s Surveillance Network: Who Can Search Your Movements?

Begin with a camera network and the question of which accounts can search its records.

Open dossier
  1. Same mechanism

    Both make privately owned cameras searchable by police. Flock does it through sharing permissions between agencies; Ring did it through a request tool Amazon said in 2024 it would stop.

    Evidence and limits
    • Flock: participating agencies can see data another customer chooses to share.
    • Amazon: in 2024 Ring said it would stop letting police request footage directly through Neighbors.

    Limit: Neither file shows Ring and Flock operating as one system.

    File 02Amazon’s Secret Surveillance Deals: Ring, Rekognition & Government Contracts
  2. Same mechanism

    The company collects first and government reaches the data later through contracts and procurement, not direct collection.

    Evidence and limits
    • Data broker: the collection point moved to private markets before any agency appears.
    • Amazon: one company sells consumer cameras, machine vision and government cloud at once.

    Limit: The Amazon file says the public record does not prove an integrated domestic surveillance system.

    File 03Data Broker Surveillance State: How The U.S. Built A Warrant-Free Tracking Market
  3. Same mechanism

    Two routes around limits on direct collection: buying commercial data, or receiving what an allied agency collected.

    Evidence and limits
    • Data broker: the warrant loophole has not been closed and reform bills have stalled.
    • Five Eyes at Home: collection, access and accountability can sit in different countries.

    Limit: Neither file claims every purchase or every exchange is unlawful.

    File 04Five Eyes at Home: How Intelligence Sharing Bypassed National Privacy Limits
  4. Documented link

    Both turn on Section 702 data the government already holds, and on what machine-scale analysis changes about searching it.

    Evidence and limits
    • Five Eyes at Home: FISA and Section 702 set the formal rules for foreign intelligence.
    • Pentagon-Anthropic: the surveillance gap is AI searching data held under Section 702 and other authorities.

    Limit: The Pentagon file does not support saying Claude was deployed for Section 702 searches.

    File 05Inside the Pentagon–Anthropic Standoff Over AI Mass Surveillance
Trail 02 / Surveillance history

How did the listening begin?

From wartime codebreaking to the 2013 leaks: one alliance, four generations of interception.

Case file / 01

1940s — The Intelligence State Is Born

Begin with the postwar settlement that made intelligence a permanent part of the state.

Open dossier
  1. Editorial pattern

    Wartime intelligence machinery was kept and formalised in 1947; the Five Eyes file traces the wartime signals cooperation that hardened into a standing partnership.

    Evidence and limits
    • 1940s: the National Security Act of 1947 is the hinge of the postwar system.
    • Five Eyes: wartime SIGINT cooperation between Britain and the US was codified as UKUSA.

    Limit: The 1940s file says later surveillance expansions were not already complete by 1949.

    File 02The Five Eyes Alliance – One Network to Watch Them All
  2. Documented link

    ECHELON is the interception system associated with that partnership, examined by the European Parliament.

    Evidence and limits
    • ECHELON: the UKUSA/Five Eyes partnership is officially documented.
    • ECHELON: the European Parliament formed a committee and passed a resolution on the system.

    Limit: The ECHELON file says documented existence does not make every dramatic claim true.

    File 03ECHELON Unveiled: Inside the Secret Global Listening Program You Were Never Meant to Know About
  3. Editorial pattern

    From intercepting international traffic to fusing databases for prediction: the same ambition of searching everything, at a different layer.

    Evidence and limits
    • ECHELON: high-volume international communications filtered through shared selection systems.
    • TIA: integrated databases and pattern recognition after September 11.

    Limit: TIA file: it influenced the surveillance imagination; later systems need their own proof.

    File 04Total Information Awareness: The Precursor to Today’s Surveillance State
  4. Editorial pattern

    Both put data held by companies at the centre of intelligence collection.

    Evidence and limits
    • TIA: commercial databases could all become inputs to a prediction system.
    • PRISM: provider-based Section 702 collection through legal demands on companies.

    Limit: Neither file says PRISM descends from TIA.

    File 05PRISM Exposed: The Surveillance Program That Changed Everything
  5. Documented link

    PRISM became public in June 2013 through Snowden's disclosures; his file places PRISM as one visible doorway among several programmes.

    Evidence and limits
    • PRISM: entered public view through reporting based on Snowden's documents.
    • Snowden: Section 215 and Section 702, with PRISM as one doorway.

    Limit: The PRISM file says the 'direct access' claim is disputed; oversight material points to compelled production.

    File 06Edward Snowden: The Exile Who Exposed the Empire
Trail 03 / AI control

How much autonomy is too much?

Five files that move from an AI test that reached real companies to the question of who keeps control at the limit.

Case file / 01

When AI Testing Reaches Real Companies: Who Controls the Boundary?

Begin with a safety test that touched companies that never agreed to take part.

Open dossier
  1. Documented link

    The testing file separates accidental reach from the deliberate misuse in the cyber-offense file: same technology, opposite failure.

    Evidence and limits
    • Testing: this is a testing-control story, distinct from deliberate misuse in AI and Cyber-Offense.
    • Cyber-offense: a state-linked group used an AI model to run most of an intrusion campaign.

    Limit: Cyber-offense claims are largely the company's own account, not an independent audit.

    File 02AI and Cyber-Offense: When the Hacker Is a Machine
  2. Same mechanism

    Both files follow the same thread: how far humans can be taken out of the loop, first running an attack, then selecting targets.

    Evidence and limits
    • Cyber-offense: automation reaches the last place a human used to be required.
    • Weapons: policy on paper asks for human judgment; reported systems narrow it.

    Limit: The weapons file does not settle whether any specific strike was lawful.

    File 03AI and Autonomous Weapons: Who Pulls the Trigger When the Machine Decides?
  3. Same mechanism

    From human control over targeting to human control over the highest-stakes decision, where time to challenge a machine's assessment is the issue.

    Evidence and limits
    • Weapons: the human-control question and where it erodes in practice.
    • Nuclear: can a person inspect and challenge the evidence before the window closes?

    Limit: Nuclear file: no public evidence that AI systems independently control nuclear launch.

    File 04AI, Nuclear Risk and the Shrinking Decision Window
  4. Editorial pattern

    The same control question at the scale of the labs: how much autonomy to grant, and who decides to scale, pause or deploy.

    Evidence and limits
    • Nuclear: faster processing can become earlier commitment.
    • Superintelligence: OpenAI says people still decide whether to scale, pause or deploy.

    Limit: The superintelligence file separates today's models, assessed as lower risk, from future systems.

    File 05Can We Control AI That Outthinks Us?
Trail 04 / Influence

Who shapes what you believe?

From a documented drugging to the modern moderation debate. The early files are documented history; the later ones are contested.

Case file / 01

Nine Days: The Unresolved Death of Frank Olson

Begin with what is documented about one scientist's last nine days.

Open dossier
  1. Documented link

    Olson was dosed with LSD at a retreat tied to the programme the MK-Ultra file documents.

    Evidence and limits
    • Nine Days: CIA chemist Sidney Gottlieb covertly dosed Olson in 1953 at an MKUltra-linked retreat.
    • MK-Ultra: a CIA programme authorised in 1953 to study altering human behaviour.

    Limit: How Olson died remains disputed; the dosing is documented.

    File 02MK-Ultra Exposed: Inside the CIA’s Mind-Control Experiments
  2. Documented link

    Both CIA Cold War programmes reached official view through the Church Committee era inquiries of the 1970s.

    Evidence and limits
    • MK-Ultra: 1975 Church Committee-era inquiries confirmed the programme.
    • Mockingbird: the Church Committee moved the press question from rumour into an oversight frame.

    Limit: The Mockingbird file says the scale, naming and command structure of the press story are less clean than the legend.

    File 03Operation Mockingbird: The CIA, the Press, and the Hidden Architecture of Influence
  3. Editorial pattern

    The Mockingbird file leaves open how far platform-era information management descends from Cold War influence architecture.

    Evidence and limits
    • Mockingbird: unresolved how far modern messaging and platform management descend from the Cold War model.
    • Fact-checkers: funding ties between verification bodies, platforms and public grants.

    Limit: The fact-checkers file says funding dependence is a conflict risk, not proof of corruption.

    File 04Fact-Checkers Funding: The Hidden Truth Behind “Independent” Verification
  4. Same mechanism

    Both start from the post-2016 push to counter online misinformation, one through funded verification, the other through government and platform channels.

    Evidence and limits
    • Fact-checkers: platform partnerships expanded after the 2016 US election.
    • Censorship complex: an intervention environment took shape after 2016 and hardened during COVID.

    Limit: The censorship file says there is no single hidden office and rejects that version of the story.

    File 05The Censorship-Industrial Complex
Trail 05 / Black budget

Who builds the state's secret tools?

Three routes by which the national-security state gets technology outside normal channels.

Case file / 01

Operation Paperclip: How the U.S. Recruited Nazi Scientists After World War II

Begin with the postwar programme that bent its own rules to acquire capability.

Open dossier
  1. Same mechanism

    Both describe a permanent national-security exception: rules bent for capability, and an agency built to outrun ordinary bureaucracy.

    Evidence and limits
    • Paperclip: an exception system that preserved capability by bending declared principle.
    • DARPA: created after Sputnik as a permanent exception to normal government process.

    Limit: Neither file claims DARPA descends from Paperclip.

    File 02DARPA: The Militarisation of Innovation
  2. Same mechanism

    Both exist to move faster than the government's own acquisition system: DARPA through research bets, In-Q-Tel through venture-style investment.

    Evidence and limits
    • DARPA: designed to create technological surprise.
    • In-Q-Tel: the interface between the CIA and venture-speed technology.

    Limit: The In-Q-Tel file stresses it operates openly; neither file claims secret control.

    File 03In-Q-Tel: How The CIA Funds Tomorrow’s Surveillance
Trail 06 / Unresolved

Why were these clusters never reviewed as a whole?

Three clusters of harm around sensitive work. Each file rejects the simple conspiracy version; each leaves the pattern unexamined.

Case file / 01

The Marconi Scientists: 25 Deaths, No Inquiry, No Answers

Begin with the 1980s deaths that reached Parliament and stalled there.

Open dossier
  1. Editorial pattern

    Both are clusters among people near sensitive work, with individual rulings on record and no review of the pattern as a whole.

    Evidence and limits
    • Marconi: questioned in Parliament in 1988; the ministerial answer on an inquiry was no.
    • Microbiologists: no aggregate inquiry into the pattern found in the update check.

    Limit: Both files say the evidence does not prove a common perpetrator, and the headline counts are disputed.

    File 02The Dead Microbiologists: 40 Researchers, One Crisis, No Answers
  2. Editorial pattern

    Another cluster among national-security personnel, with a cause that remains unsettled. It concerns injury, not deaths.

    Evidence and limits
    • Microbiologists: a real cluster, not a solved case.
    • Havana: reported health incidents since 2016; public assessments have not linked a foreign actor.

    Limit: The Havana file says the weapon hypothesis is not fantasy and is not proven.

    File 03Havana Syndrome: The Weapon That Officially Doesn’t Exist
Trail 07 / Alien files

Where does the real secret end and the legend begin?

A documented secret base, and a second site that rests on witness accounts.

Case file / 01

Area 51: Black-Budget Aerospace Secrecy

Begin with the base that is real, and how the legend grew on top of it.

Open dossier
  1. Documented link

    The S-4 claim borrows credibility from Groom Lake's documented secrecy. The S-4 file tests what traces a real hidden facility should leave.

    Evidence and limits
    • Area 51: the Groom Lake site is documented; the S-4 layer entered mass culture in 1989.
    • S-4: no authenticated public record currently names S-4 as a site.

    Limit: Area 51 file: the aerospace record does not authenticate recovered craft.

    File 02The S-4 Problem: Why Nothing Leaks
Between the trails

Where the trails cross

Some files sit on more than one route. These are the crossing points.