AI Manipulation Detection Software
Budget: $250 – $750 USD
I’m building a cross-platform application that can inspect user-supplied files—video, audio, photos, and written documents—and tell the user whether any part of that asset has been generated or altered by AI. The priorities are clear:
• Flag deepfake content in video frames.
• Detect voice synthesis in audio tracks.
• Identify image alteration in still photos.
• Spot AI-generated or AI-edited passages in text files.
A single, intuitive interface should let someone drop a file in, trigger the scan, and instantly see a confidence score plus a short explanation of the features that raised the alert. Under the hood I expect you to combine established media forensics with modern machine-learning techniques; think OpenCV or FFmpeg for data extraction, TensorFlow/PyTorch for model inference, and an NLP classifier for written text detection.
Deliverables
- Source code with clear build/run instructions.
- Pre-trained models (or training scripts) capable of the four checks above.
- A minimal GUI or command-line tool that ties everything together.
- A short report outlining the datasets used, evaluation metrics, and accuracy results on a sample set I will provide.
The solution must run offline after installation and should be modular enough that future manipulation types can be slotted in without rewriting the core. If you’ve tackled media forensics, deepfake detection, or AI-generated text analysis before, I’d love to see examples of your work along with your proposal.
Confidentiality Notice
This project involves sensitive information protected under applicable privacy laws such as the Privacy Act of 1974 and the General Data Protection Regulation (GDPR). By bidding on this project, freelancers acknowledge their understanding of these confidentiality requirements.
Before being awarded this project, the selected freelancer will be required to execute a formal Non-Disclosure Agreement (NDA) and any additional legally binding agreements necessary to ensure compliance with confidentiality standards.
• Flag deepfake content in video frames.
• Detect voice synthesis in audio tracks.
• Identify image alteration in still photos.
• Spot AI-generated or AI-edited passages in text files.
A single, intuitive interface should let someone drop a file in, trigger the scan, and instantly see a confidence score plus a short explanation of the features that raised the alert. Under the hood I expect you to combine established media forensics with modern machine-learning techniques; think OpenCV or FFmpeg for data extraction, TensorFlow/PyTorch for model inference, and an NLP classifier for written text detection.
Deliverables
- Source code with clear build/run instructions.
- Pre-trained models (or training scripts) capable of the four checks above.
- A minimal GUI or command-line tool that ties everything together.
- A short report outlining the datasets used, evaluation metrics, and accuracy results on a sample set I will provide.
The solution must run offline after installation and should be modular enough that future manipulation types can be slotted in without rewriting the core. If you’ve tackled media forensics, deepfake detection, or AI-generated text analysis before, I’d love to see examples of your work along with your proposal.
Confidentiality Notice
This project involves sensitive information protected under applicable privacy laws such as the Privacy Act of 1974 and the General Data Protection Regulation (GDPR). By bidding on this project, freelancers acknowledge their understanding of these confidentiality requirements.
Before being awarded this project, the selected freelancer will be required to execute a formal Non-Disclosure Agreement (NDA) and any additional legally binding agreements necessary to ensure compliance with confidentiality standards.