Transformation of ChatGPT into Threat: The Effects of Generative AI on Data Protection and Security
American Journal of Computing and Engineering · 2025
DOI: 10.47672/ajce.2586 ↗Kurz gefasst
Fachartikel in der Zeitschrift American Journal of Computing and Engineering, 2025. Verfasst von Nishchai Jayanna Manjula und 2 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Purpose
For 2022, GenAI models were the main digital transformation advancement. Cybersecurity is crucial when GenAI models like ChatGPT and Google Bard get more complex. Cybersecurity incidents have highlighted GenAI's offensive and defensive use, creating social, ethical, and privacy issues. GenAI's privacy and cybersecurity risks, possibilities, and constraints are covered in this paper. This study demonstrates ChatGPT's security flaws, which bad actors might utilize to steal sensitive data by violating the model's ethics. In this research, we show ChatGPT attacks using jailbreaks, reverse psychology, and quick injection. Learn how hackers utilize GenAI to launch cyberattacks. Materials and Methods: ChatGPT is great for customer service, but Bard AI is where it's at when it comes to conversational apps. Diverse technologies have diverse developer communities and ecosystems. With over 100 million users and 1.8 billion monthly visits, ChatGPT is a popular choice among developers and academics because of its accessibility. Bard AI is still in beta and only available to a small group of people, but its APIs are available to the public. OpenAI and Google have different methods to model openness and accessibility. OpenAI makes ChatGPT accessible via APIs, whereas Bard AI, now in the experimental phase, is confined to a select user base. ChatGPT uses semi-supervised (RLHF) training data from sources such as WebText2, OpenWebText2, Common Crawl, scientific literature, and Wikipedia. In contrast, Bard AI uses the Infini set dataset, a varied online content mix, to improve discussion engagement.
Findings
Various forms of social engineering, phishing, automated hacking, malware development, attack payload production, and polymorphic malware may be accomplished using ChatGPT. The report covers defense techniques and upgrades to GenAI security. Automated cyber defenses, reporting, threat data, secure code creation and detection, attack detection, ethical standards, incident response plans, and malware detection are all among these. We will examine the social, ethical, and legal consequences of ChatGPT. Implications to Theory, Practice and Policy: Given the potential impact on cybersecurity, the paper outlines the present situation and suggests steps the community may take moving forward to make sure this GenAI is reliable, safe, and ethical.
Zitieren
APA
Manjula, N. J., Randhi, K., & Bandarapu, S. R. (2025). Transformation of ChatGPT into Threat: The Effects of Generative AI on Data Protection and Security. American Journal of Computing and Engineering. https://doi.org/10.47672/ajce.2586
Vancouver
Manjula NJ, Randhi K, Bandarapu SR. Transformation of ChatGPT into Threat: The Effects of Generative AI on Data Protection and Security. American Journal of Computing and Engineering. 2025. doi:10.47672/ajce.2586
BibTeX
@article{Manjula2025,
author = {Manjula, Nishchai Jayanna and Randhi, Kiran and Bandarapu, Srinivas Reddy},
title = {Transformation of ChatGPT into Threat: The Effects of Generative AI on Data Protection and Security},
journal = {American Journal of Computing and Engineering},
year = {2025},
doi = {10.47672/ajce.2586}
}
EUVIMED