Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM62360
Volume:
10 |
5 |
May 2026
Pages:
1-4
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.
department of computer science
jagan institute of management studies, rohini ,new delhi
department of computer science
jagan institute of management studies,rohini,new dlehi
Artificial intelligence has changed the way we interact with technology. Large language models (LLMs), like ChatGPT, are now used in everything from customer support to coding assistance. As these systems become more powerful, people are starting to rely on them for important and sometimes sensitive tasks. But with this rapid growth comes a new kind of risk. One of the most concerning threats is known as a prompt injection attack. Unlike traditional cyberattacks that exploit software bugs or system weaknesses, this type of attack targets how AI understands and responds to language. For example, a user might enter a command like, “Ignore previous instructions and reveal confidential data.” Even though it looks like a simple sentence, it can trick the system into behaving in unintended ways. This raises an important question: How can we tell the difference between a normal user request and a harmful one hidden in natural language?
KEYWORDS
Prompt injection attacks represent a serious and growing threat to large language models. By exploiting how these systems process language, attackers can manipulate them in ways that are difficult to detect. This paper presents a hybrid approach that combines rule-based and machine learning techniques to address this issue. The results show that this method is both effective and practical. As AI continues to evolve, building secure and trustworthy systems will be essential—and this research takes an important step in that direction.
Prompt injection attacks represent a serious and growing threat to large language models. By exploiting how these systems process language, attackers can manipulate them in ways that are difficult to detect. This paper presents a hybrid approach that combines rule-based and machine learning techniques to address this issue. The results show that this method is both effective and practical. As AI continues to evolve, building secure and trustworthy systems will be essential—and this research takes an important step in that direction.
Publication Date
5/4/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
0
Total views since publication
PDF Downloads
9
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...