01
AI-generated images and face manipulation
Review signs associated with synthetic images, face swaps, inpainting, splicing and other forms of digital alteration.
Digital Evidence Verification
Deepfake & AI-Generated Media Detection
Examine disputed images, video and audio for indicators of synthetic generation or manipulation, with the limits stated clearly.
The Brief
A suspicious image, recording or video rarely exists in isolation. CPIS reviews the media together with its source, metadata and surrounding evidence. Voice-clone analysis receives the greatest caution because audio alone may not support a reliable verdict.
Who this is for
Built for sensitive,
high-stakes situations.
- 01Businesses responding to executive, employee or brand impersonation
- 02Law firms and insurers reviewing disputed digital evidence
- 03HR and compliance teams investigating suspected media manipulation
- 04Individuals facing scams, harassment, blackmail or reputational harm involving synthetic media
Confidential enquiry
Discuss your case
Complete the form below to enquire about Deepfake & AI-Generated Media Detection. A CPIS case handler will contact you confidentially.
Media examined
More than a detector score.
02
Deepfake video and impersonation
Examine visual continuity, lip synchronisation, compression, source history and other indicators relevant to manipulated video.
03
AI voice and audio concerns
Assess suitable recordings and the surrounding communication trail. Audio detection is less reliable and more sensitive to duration, compression, background noise and the channel through which it was recorded.
04
Context and source verification
Compare the media with original files, account activity, publication history and other evidence that may support or contradict authenticity.
Methodology
Discreet. Lawful. Evidence-driven.
- 01
Preserve the best available source
We identify and retain the original or highest-quality file available, document its source and calculate hashes where appropriate.
- 02
Metadata and provenance review
File properties, editing history, compression and available source information are examined before interpreting detector output.
- 03
Current, layered examination
Case-appropriate visual, audio and contextual methods are assessed together. Significant tools and versions are recorded, and no single automated detector is treated as definitive proof.
- 04
Confidence-based reporting
Findings distinguish observed indicators from conclusions, state limitations and alternative explanations, and identify when the material is inconclusive or requires a separately qualified expert scope.
Deliverables
Findings with context.
- 01Schedule of the files and sources examined
- 02Documented preservation steps and file hashes, where appropriate
- 03Methods, significant tool versions and authenticity indicators
- 04Confidence-based findings with limitations and alternative explanations
- 05A clear statement when the evidence does not support a reliable verdict
- 06Supporting exhibits and a confidential written report
The highest-caution medium
Audio findings may remain inconclusive.
Voice-clone detectors can be affected by new generation models, short samples, telephone channels, compression and background noise. CPIS will not label a recording authentic or synthetic from a detector score alone. The report states the strength of the evidence and should not be treated as the sole basis for a court, HR or personal decision.
Questions, answered.
Speak To An Investigator
When the matter is sensitive, the firm you choose matters most.
Speak directly with a licensed investigator. Free, fully confidential, no obligation, most matters can be assessed in a single call.
- Office
- 197A Thomson Road, Goldhill Centre, Singapore 307635
- Response
- Same-day for time-sensitive matters. 24 to 48 hours standard.
- Discretion
- Anonymous intake available. Your enquiry stays private.