Synthetic Media, Deepfakes, and Digital Trust

An evidence-based account of what fails, what holds, and how each claim can be checked

Contents

  1. 1. Foundations of Synthetic Media Generation

    GANs, diffusion models, voice cloning, and video generation — and why the engineering details decide which defences can possibly work.

    13 min · 2,934 words

  2. 2. A Taxonomy of AI-Generated Disinformation Campaigns

    With election case studies, and an argument that the decisive variable is not the artefact but the window in which it cannot be answered.

    11 min · 2,410 words

  3. 3. Deepfakes as a Cybersecurity Threat Vector

    Fraud, identity theft and social engineering. The documented record contains no case in which detecting the fake is what saved the money.

    13 min · 2,822 words

  4. 4. Deep Learning and Multimodal Deepfake Detection

    What detectors actually learn, why benchmark scores do not survive contact with unfamiliar content, and the arithmetic that decides whether a good detector is useful to you.

    14 min · 3,018 words

  5. 5. Voice Cloning and Audio Deepfake Detection

    Three seconds is enough to clone a voice. The channel that carries nearly every real attack is also the channel that destroys the evidence a detector needs.

    12 min · 2,651 words

  6. 6. Benchmarking, Datasets and Cross-Domain Generalisation

    Detectors expire for reasons that have nothing to do with attackers, and the benchmarks that certify them cannot see it happening.

    12 min · 2,591 words

  7. 7. Adversarial Robustness and Evasion

    Flipping the lowest bit of every pixel turns a 0.95 AUC detector into one that is reliably wrong. And if that happened to you, nothing would tell you.

    12 min · 2,595 words

  8. 8. LLM-Generated and Multimodal Disinformation

    Text detectors do not measure provenance. They measure fluency, and that is why their errors land on the same people every time.

    12 min · 2,739 words

  9. 9. Emerging Generative Threats

    Real-time cloning, text-to-video, agentic impersonation and synthetic identities. A stress test of this book’s own recommendation against the capabilities arriving next.

    14 min · 3,019 words

  10. 10. Content Provenance and Watermarking

    Provenance asks a bounded question about material you control, instead of an unbounded one about material an adversary controls. It becomes useful only at the point where absence starts to mean something.

    15 min · 3,223 words

  11. 11. The Legal and Regulatory Landscape

    Transparency duties, likeness rights and evidence. Law has moved fastest on the question with the least purchase and slowest on the one with the greatest institutional exposure.

    13 min · 2,919 words

  12. 12. Building an Organisational Defence Framework

    Every control here is traced to the claim that justifies it, and every metric is one you can compute without knowing whether you have been attacked.

    14 min · 3,012 words

Every substantial factual claim carries an identifier resolving to the evidence ledger, which records the source, its status, and what that source does not establish. 33,933 words, 12 chapters, 47 claims. Free to read, no account required.