To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
A Human-Authored or GenAI-Generated: Who is Creating the Content |
|
Author |
Jalal S Alowibdi |
| Citation |
Vol. 25 No. 6 pp. 128-134 |
|
Abstract |
Online Social Networks (OSNs) share a huge volume of content data. Nowaday, user-generated content in OSNs presents challenges for showing their authenticity and origin. In particular, user-generated contents in tweets from X (formally Twitter) have shown difficulties in knowing its authenticity and origin. With the new regulations and rules for the tweets content in Twitter from countries and Twitters itself, people are trying to generate more content than it was before. The huge amount of data generated from tweets' contents are worth an investigation to know their authenticity and origin. From commercial endeavors to law enforcement and social dynamics, understanding the authenticity and origin becomes more important. Thus, in our study we are exploring the task of predicting the source of tweets, with a particular focus on distinguishing between human-authored and GenAI-generated content. Through our analysis and experimentation, we have achieved a remarkable 94% accuracy in attributing tweets to their original source. Additionally, we have identified and listed various distinguishing characteristics between GenAI-generated tweets and human-authored tweets, further enhancing our understanding of the differences in their content and structure. |
|
Keywords |
Generative AI; Artificial Intelligence; Linguistic Patterns; Text Classification; GenAI-generated; Human authored |
|
URL |
