Artificial Intelligence in Detecting Fake Academic Certificates
Sahil Gole1, Abhishek Bhosale 2, Dr. Swati Joshi3
1,2 Department of Computer Science, PVG’s College of Science and Commerce
3 Research Guide, Department of Computer Science, PVG’s College of Science and Commerce
Abstract
The selling of fake academic diplomas has become a major international business that endangers the usual legitimacy of academic institutions and the security of the labor force. With billions of dollars in illegal revenues through the operation of the so-called diploma mills and the digital methods of forgery actually moving beyond the reach of the naked eye, the old methods of manual verification, usually marked by bureaucratic latency and human error, have become obsolete. The proposed research report is a thorough study exploring how Artificial Intelligence (AI) can be
radically applied to the field of document forensics and credential validation. We discuss a multilayered technological model based on Deep Learning (DL) networks, i.e., Convolutional Neural Networks (CNNs) and Graph Convolutional networks (GCNs) and forensic tools such as Error
Level Analysis (ELA) and Optical Character Recognition (OCR). We go even further to the application of blockchain technology as an unchangeable register to support the authenticity of the digital credentials. This report reveals the promise of using AI in automating the detection of micro-anomalies that can be used to identify forgery by conducting a stringent analysis of methodologies, such as the mathematical foundations of the Structural Similarity Index Measure (SSIM) and the cryptographic logic of smart contracts. We offer comprehensive performance measures of the modern literature where accuracy rates are above 97 percent in the controlled settings, yet we approach the so-called innovative threat to the adversarial attacks with seriousness and the ethical discussions on transparency of algorithms are also stated. This paper seeks to offer a conclusive technical architecture of the implementation of AI-based verification systems.
Keywords:
The AI, artificial intelligence, Fake Certificate Detection, Deep learning, Computer vision, Pentium blockchain, Convolutional neural network, error level analysis, Optical Character Recognition, academic integrity, Graph convolutional networks.
