Upload the video above as Step 1 of your verification workflow. Get an AI probability score in seconds, then continue with the steps below.
Verifying whether a video is authentic has never been more important — or more challenging. This step-by-step guide covers every technique available in 2026 for confirming that video footage is genuine and unmanipulated.
Step 1: Run an AI Detection Tool
Upload the video to our free Sora AI Detector for an instant AI-generation probability score. This is your fastest first filter. For a full explanation of how detection works, read our complete detection guide.
Step 2: Inspect Metadata
Use a free tool like ExifTool to inspect the file’s embedded metadata. Authentic camera footage contains camera model, lens data, timestamps, and often GPS. AI-generated video typically has none of this. Note: metadata can be added or stripped, so treat absence as a flag, not proof.
Step 3: Reverse Video Search
Upload key frames to Google Reverse Image Search or InVID/WeVerify. If the event shown is real, other footage, witnesses, and news reports will exist. Total absence of corroboration is a major red flag.
Step 4: Visual Frame Inspection
Step through the video frame by frame. Look for the 10 signs of AI-generated video: finger count errors, background flickering, physics anomalies, texture uniformity, and lip-sync issues.
Step 5: Check Audio
Does the audio match the environment? Is the voice consistent with the speaker’s known recordings? AI voice synthesis can be convincing but often has subtle cadence and ambient sound mismatches.
Step 6: Source Interrogation
Who first posted this video? Is it a verified account with a traceable history? Anonymous accounts posting dramatic footage with no corroboration are your primary red flag. For newsroom-specific protocols, see our journalist verification guide and for legal contexts read our AI video legal evidence guide.
Building a Personal Verification Habit
The most effective protection against AI video misinformation is not any single tool — it is a habitual scepticism combined with a systematic verification approach. The questions to ask before trusting any dramatic video: Who first posted this? When was the account created? Are there other independent recordings of this event? Does the physics look right? Does the detail look right? Has any credible news organisation reported on this event independently?
These questions take 60 seconds to work through and catch the vast majority of AI video disinformation at the point where most people encounter it — social media feeds before independent fact-checking organisations have had time to respond.
Verification Tools Reference
- Sora AI Detector: Free, instant AI generation and deepfake detection
- ExifTool: Free, cross-platform metadata inspection tool
- InVID/WeVerify: Browser extension for reverse video search and frame extraction
- Google Reverse Image Search: Frame-level reverse search for identifying original footage sources
- contentcredentials.org: C2PA metadata verification tool from the Content Authenticity Initiative
For context on the broader AI video threat, read our AI video misinformation guide and our AI video fraud cases roundup. Stay current with our AI News section. For professional verification contexts, see our guides for journalists and legal professionals.