..
AI-Enhanced Blockchain
Data Vault for Secure and
Tamper-Proof Storage
Group Member name: Supervisor name:
Tanishka Chauhan - 221010250 Date:
Date: 04-03-2025
Debashish Padhy - 221020424 Dr. Srinivasa KG &
Akshat Srivastava-221020402 Dr. Malikharjuna Rao K
Dr. Shyama Prasad Mukherjee
International Institute of Information
Technology, Naya Raipur
International Institute of Information Technology, Naya
Content
• Introduction
• Motivation: Issues and Challenges
• Literature Review
• Problem definition
• Objectives
• Proposed Framework/Model/System/Methodology
• Explain with a block diagram and others
• Experimental results and discussion
• Plan of action for the project
• Conclusion and Future Directions
* International Institute of Information Technology, Naya Raipur 2
Introduction
• Data Security Challenges: Traditional
accounting systems are vulnerable to
tampering and unauthorized access, requiring
a more secure solution.
• Blockchain & AI for Protection: Blockchain
ensures immutable data storage, while AI
detects anomalies and suspicious activities in
real time.
• Implementation Challenges: High costs,
scalability issues, and integration complexity
make Blockchain adoption difficult for AIS.
Source: Appinventiv
* International Institute of Information Technology, Naya Raipur 3
Introduction
Accounting Information Enterprise Resource Blockchain and AI for
Systems (AIS) Planning (ERP) AIS & ERP
AIS collects, stores, and ERP systems integrate
Blockchain ensures
processes financial data multiple business
immutable financial
to generate reports for functions, including AIS,
records, while AI detects
decision-making. into a centralized system.
anomalies, creating a
Integrated with ERP, AIS They improve data
secure and intelligent AIS
enhances financial sharing, automation, and
& ERP system.
accuracy and prevents efficiency across an
fraud. organization.
4
Motivation
• Data Security Risks: Traditional databases are like
weakly locked houses, vulnerable to fraud and
data breaches.
• Blockchain for Protection: A tamper-proof diary
where records can’t be changed, ensuring data
security.
• AI for Smart Monitoring: Acts like a security
camera, detecting unusual activity and preventing
fraud.
Source: Comparitech
* International Institute of Information Technology, Naya Raipur 5
Literature Review
[Link]. Reference Technique Used Database Used Accuracy/ Remarks
Measures (Observations/Limitations)
1. Sarwar, M.I., Iqbal, M.W., Alyas, T., Conceptual Relational Database N/A Proposes a cost-effective framework to ensure AIS
Namoun, A., Alrehaili, A., Tufail, A., & framework, Management (Conceptual) data integrity using Data Vaults backed by Blockchain.
Tabassum, N. (2021). Data Vaults for cryptographic Systems (RDBMS) Aims to prevent alterations and identify data
Blockchain-Empowered Accounting primitives, SHA256, with Blockchain breaches. Describes applying SHA256 on data inside
Information Systems. IEEE Access, 9, data vault the block and then chaining that block to secure data
117306-117315. architecture vaults. Does not implement or evaluate the
performance of the proposed system.
2. Fullana, O., & Ruiz, J. (2019). Literature review, N/A N/A Analyzes the advantages of blockchain technology in
Accounting information systems in the analysis of AIS, highlighting issues like governance, transparency,
blockchain era. advantages and continuous audit, smart contracts, and the roles of
disadvantages accountants/auditors. It also examines early reactions
from the accounting industry and regulators.
3. Demi, S., & Haddara, M. (2018). Do Case study, Case study on Qualitative Suggests cloud-ERP systems may have longer
Cloud ERP Systems Retire? An ERP Literature Review, Xledger (a cloud- analysis of lifecycles overall but shorter retirement phases
Lifecycle Perspective. Procedia ERP Lifecycle Model based ERP system ERP compared to on-premise systems. Limited empirical
Computer Science, 138, 587–594. vendor in Norway) lifecycle, validation beyond a single case study.
focusing on
the
retirement
phase
4. Chase, B. & MacBrough, E. (2018). Algorithmic XRP Ledger Conditions Focuses on theoretical analysis of XRP LCP. Does not
Analysis of the XRP Ledger Consensus explanation, formal (abstractly treated) for safety & address practical implementation challenges,
Protocol. Ripple Research. analysis, network liveness performance evaluation, or impact of network
modelling based on conditions. Mentions a more general fault model
UNL overlap; requires >90% UNL overlap. Serves primarily to show
Corrected that the XRP Ledger is safe in the interim while
* International Institute of Information UNL overlap
Technology, Nayatransitioning
Raipur to Cobalt 6
bound
Problem Definition/Problem statement
• Financial Data at Risk: Traditional accounting systems can be altered or hacked,
leading to fraud, errors, and trust issues in financial reporting.
• Lack of Smart Security: Existing security measures fail to track unauthorized
changes in real time, making it hard to detect data manipulation.
• Challenges in Blockchain Adoption: While Blockchain provides tamper-proof
storage, high costs and integration difficulties slow down adoption in accounting
systems.
• AI + Blockchain Solution: A Blockchain-based Data Vault with AI-powered threat
detection ensures secure, unchangeable records while detecting fraud in real time.
* International Institute of Information Technology, Naya Raipur 7
Objective (s) of the Project
• Protect Data with Blockchain : Use Blockchain to store financial records securely and
permanently, preventing tampering or unauthorized changes.
• Detect Threats with AI : AI monitors user activity in real-time, spotting and stopping
suspicious access before damage occurs.
• Build a Secure Data Vault : Store sensitive data in an encrypted, Blockchain-backed
system to keep it safe from cyberattacks.
• Make Blockchain Faster & Cheaper : Design a lightweight and cost-effective
Blockchain model that ensures security without high processing costs.
* International Institute of Information Technology, Naya Raipur 8
Proposed Framework/Methodology
Data Vault
● Function: Securely stores financial and transactional data.
● Key Feature: Encrypted, structured storage.
● Tech Stack: MySQL, PostgreSQL, AES Encryption.
Blockchain Layer
● Function: Ensures data integrity using cryptographic hashes.
● Key Feature: Immutable records, tamper-proof data.
● Tech Stack: Ethereum, Hyperledger Fabric.
AI-Based Anomaly Detection
● Function: Identifies suspicious activities from user logs.
● Key Feature: Machine learning-powered threat detection.
● Tech Stack: TensorFlow, Scikit-learn, Federated Learning.
* International Institute of Information Technology, Naya Raipur 9
Proposed Framework/Methodology
User Interface
● Function: Monitors access logs and security events.
● Key Feature: Displays real-time security insights.
● Tech Stack: [Link], [Link], Dashboard APIs.
Verification Process
● Function: Confirms data authenticity using hash comparisons.
● Key Feature: Detects tampering attempts.
● Tech Stack: SHA-256
Automated Alerts & Response Mechanism
● Function: Sends notifications and locks compromised
accounts.
● Key Feature: Real-time security alerts.
● Tech Stack: Webhooks, SMTP, Authentication APIs.
* International Institute of Information Technology, Naya Raipur 10
Proposed Framework/Methodology
Component Technology/Tools
Data Vault MySQL/PostgreSQL (database), HashiCorp
Vault (cryptography library for encryption).
Blockchain Layer Ethereum (public/consortium), Hyperledger
Fabric (private), IPFS (lightweight storage).
Smart Contracts Solidity (to automate hash storage/retrieval).
AI Models Isolation Forest, Autoencoders, Random
Forest/XGBoost, LSTM models.
AI Libraries Scikit-learn, TensorFlow, PyTorch,
Pandas/NumPy for preprocessing.
Integration Layer Flask/Django (backend), REST
APIs/WebSockets for communication.
User Interface [Link]/[Link] (frontend), [Link]/[Link] for
visualizations.
11
Proposed Framework/Methodology
Step 1: Data Vault :- Sensitive data is collected and encrypted using
HashiCorp Vault or Python’s cryptography library.
Encrypted data is stored in a relational database like MySQL/PostgreSQL.
Step 2: Blockchain Layer :- Generate a cryptographic hash (e.g.,
SHA256) of each record in the database.
Store these hashes on the blockchain using smart contracts for
immutability.
Step 3: AI-Based Anomaly Detection :- The AI system monitors access
logs and user behavior in real-time.
Models like Isolation Forest or Autoencoders detect anomalies such as
unauthorized access attempts or unusual data modifications.
Step 4: Anomaly Detection Check :- The system asks: "Was an anomaly
detected?"
■ If NO → The process ends, and data is sent to the User
Interface (Step 7).
■ If YES → Proceed to Step 5.
Step 5: Verification Process :- If an anomaly is detected, a verification
process is [Link] ensures whether the change was legitimate or
tampered.
12
Proposed Framework/Methodology
Step 6: Hash Matching :- The system checks whether the hashes
match (i.e., the expected and actual data remain consistent).
○ If YES → Data is sent to the User Interface (Step 7).
○ If NO → The system triggers security actions (Step 8).
Step 7: User Interface :- If no anomalies were detected or data
integrity is verified, the user can access the data normally.
Step 8: Automated Alerts & Response Mechanism :- If hash
mismatch is found, the system activates an automated alert
mechanism.
Step 9: Admins Notified :- Administrators are notified about the
detected tampering for immediate action.
Step 10: Lock Accounts or Roll Back :- As a final security measure,
the system may lock user accounts or roll back data to a previous
secure version to prevent damage.
13
Plan of action for the remaining project work
• Complete integration between the Data Vault, Blockchain Layer, and AI
System.
• Develop a user-friendly dashboard for monitoring alerts.
• Test the system with real-world datasets (e.g., access logs from
financial or healthcare systems).
• Optimize performance (e.g., reduce latency in blockchain
transactions).
* International Institute of Information Technology, Naya Raipur 14
Conclusion and Future Directions
• Secure & Intelligent System: Combines Blockchain for tamper-proof storage and
AI for real-time anomaly detection.
• Optimized Blockchain Implementation: Use lightweight blockchain models
(IOTA/IPFS) to reduce storage and costs.
• Advanced AI Security: Implement Federated Learning & Quantum-Resistant
Encryption for enhanced protection.
• Scalability & Integration: Extend to IoT security, multi-blockchain verification, and
cross-industry applications.
* International Institute of Information Technology, Naya Raipur 15
..
Thank You
Dr. Shyama Prasad Mukherjee
International Institute of Information
Technology, Naya Raipur
* International Institute of Information Technology, Naya Raipur 16
