Divel - Digital Evidence Locker

Inspiration

With the increase in generation of AI images, videos, and audio, it has become difficult to prove the authenticity of the digital evidence since the existing evidence management solutions have only focused on storage but there are no proven means for maintaining an unbreakable chain of custody. For addressing this issue, we developed DIVEL (Digital Evidence Locker), which is a digital locker for evidences using Artificial Intelligence, Blockchain, and AWS Cloud to keep evidence authentic and encrypted through the entire process.

What it does

DIVEL is an artificial intelligence-based digital evidence management system that targets law enforcement agencies, forensic laboratories, and judiciary systems.

Our system allows investigators to:

Upload digital evidence (images, video, audio, and URL)
Detect deepfake and edited media by utilizing artificial intelligence technology
Create SHA-256 hash for unique digital fingerprints
Encrypt digital evidence with the use of AWS Key Management Service (KMS)
Store encrypted evidence on Amazon S3
Store evidence metadata on Amazon DynamoDB
Log hashes and chain of custody data on the Ethereum blockchain
Ensure evidence integrity at any stage of the investigation process

Any change in the evidence makes its hash different:

[ H(E_{original}) \neq H(E_{modified}) ]

How we built it

DIVEL integrates AI, cloud computing, and blockchain into a unified workflow.

Technology Stack Frontend: Flutter Backend: Node.js + Express.js Blockchain: Ethereum & Solidity Smart Contracts Cloud Storage: Amazon S3 Database: Amazon DynamoDB Encryption: AWS Key Management Service (KMS) API Management: Amazon API Gateway Security: SHA-256 Hashing AI: Deepfake Detection Models for Image, Video, Audio, and URL Verification Workflow Investigators upload digital evidence through the Flutter application. Requests are securely routed via Amazon API Gateway to the backend services. The backend generates a SHA-256 hash for integrity verification. AI models analyze the uploaded content for deepfakes or tampering. Evidence is encrypted using AWS KMS before storage. The encrypted files are stored in Amazon S3. Evidence metadata, investigation details, and status are stored in Amazon DynamoDB. The evidence hash, timestamp, and chain-of-custody records are written to the Ethereum blockchain, ensuring they cannot be altered.

This architecture provides secure storage, immutable auditing, and scalable evidence management.

Challenges we ran into

There were many technological issues faced in the development of DIVEL:

Incorporating AI inference with cloud-based services with low-latency Implementing encryption and key management with AWS KMS Synchronization of Amazon S3, DynamoDB, and blockchain data Reducing storage overheads by storing large evidence files in Amazon S3 but immutably hashing them on the blockchain Building a highly scalable system that can handle multimedia evidence in a secured manner

These issues were successfully resolved in order to develop a reliable end-to-end prototype for digital evidence management.

Accomplishments that we're proud of

We are proud of having created an end-to-end digital evidence management platform which successfully uses AI, Blockchain and AWS Cloud technology into one cohesive workflow.

Some of our major achievements in this project include:

Created a safe pipeline for storing the evidence using Amazon API Gateway, AWS KMS, Amazon S3 and Amazon DynamoDB.
Used blockchain technology to build chain of custody using Ethereum smart contracts and ensure that evidence cannot be altered by any other party.
Developed a way to use AI deepfake detection to check whether the evidence was manipulated before entering the process of investigation.
Ensured integrity of evidence using SHA-256 hashing and encrypted storage.
Developed a system architecture that ensures separation of functions for secure storage of files, their metadata and auditing of all actions.
Successfully created a prototype proving the concept that modern cloud services, AI and blockchain can make a difference in digital forensics.

What we are most proud of accomplishing is developing a solution to a real life problem of cybersecurity and law enforcement using multiple high-tech solutions.

What we learned

Creating DIVEL provided us with valuable experience in the development of a unified secure environment based on several different technologies.

Namely, we learned how to:

Create secure cloud architecture based on AWS services Use AWS KMS to encrypt sensitive data Manage structured metadata with Amazon DynamoDB Create RESTful APIs with Amazon API Gateway Create smart contracts on Ethereum for immutable audit trails Use cryptographic hashing of evidence Implement AI-driven deepfake detection

Most of all, we learned how AI, blockchain, and cloud services could be combined to increase confidence in digital forensics.

What's next for Divel - Digital Evidence Locker

We aim to transform DIVEL from a prototype into a full-fledged solution for managing digital evidence in law enforcement and forensic institutions.

The following actions should be taken in the near future:

Development of AI abilities for detecting deepfakes and generative AI artifacts more accurately.
Real-time analysis and validation of the evidence.
Role-based access control and multi-factor authentication for investigators, forensic specialists, and judges.
Secure evidence transfer between police departments, forensic labs, and courts, maintaining the chain of custody.
Utilizing other AWS services to increase availability, monitoring, and disaster recovery.
Creation of dashboards for case tracking, evidence analysis, and audit reports.
Pilot installations with law enforcement organizations to get user feedback.

Our ultimate goal is to create a digital evidence ecosystem that will assist in combating AI-generated misinformation and establishing trust in the legal system.