Fake AI

DE EN

Spotting synthetic text, images and voices before they cost you.

How to detect AI-generated content and deepfakes: image forensics, voice-clone scams, AI text detectors and their error rates, content credentials and company policy.

Images & video

Artefacts, metadata, provenance: checking whether a picture or clip is real.

  1. How to Tell If an Image Is AI Generated: 9 ChecksPeople spot AI faces only 48% of the time, worse than a coin flip. How to tell if an image is AI generated with provenance data, watermarks and context.
  2. How to Spot a Deepfake Video: What Still Gives It AwayThe best deepfake detector in Meta's 2020 challenge scored 65% on unseen clips. How to spot a deepfake video by source, face edges, audio and live tests.
  3. Content Credentials (C2PA): How the Photo Label WorksContent Credentials (C2PA) sign a photo's origin and edits. What the label proves, why it is usually missing and how to read it on any file in two minutes.
  4. How to Find the Original Source of an ImageTo find the original source of an image, search it in three engines, sort by date and compare resolution. Workflow, tools and the signs of a true original.

Voice & scams

Cloned voices, CEO fraud, fake support calls and how to verify a caller.

  1. AI Voice Cloning Scam: How It Works and How to Stop ItAn AI voice cloning scam needs only a few seconds of audio. How the fake family emergency call works, what the numbers say and the one habit that beats it.
  2. Family Safe Word: How to Set One Up Against AI ScamsA family safe word stops cloned-voice scam calls in seconds. How to choose one that can't be found online, share it safely and use it without giving it away.
  3. Deepfake CEO Fraud: Lessons From Arup and FerrariDeepfake CEO fraud cost Arup about US$25 million in one video call. What happened, why a Ferrari manager stopped a copycat and which payment controls work.
  4. How to Detect an AI Voice on a Call or RecordingListeners caught only 73% of AI voices in a UCL study. How to detect AI voice clones by sound, timing and content, which tools help and when to just hang up.

AI text

Detectors, watermarks, false positives and fair rules for schools and teams.

  1. Are AI Detectors Accurate? What the Studies ShowAre AI detectors accurate? Studies found false positive rates up to 61% for non-native writers. The numbers, why detectors fail and what they are fit for.
  2. Falsely Accused of Using AI? How to Prove You Wrote ItFalsely accused of using AI on an essay? Gather version history and drafts, ask for the evidence, cite detector studies and use this email template to respond.
  3. AI Text Watermark: How SynthID Works and Where It FailsAn AI text watermark hides a statistical pattern in word choice. How SynthID Text works, why OpenAI held back its own and what the EU AI Act requires by 2026.
  4. EU AI Act Article 50: AI Labelling Rules and DeadlinesEU AI Act Article 50 applies from 2 August 2026: chatbot disclosure, machine-readable marking, deepfake labels. Duties, deadlines and fines up to €15m.