The art of deception: humanizing AI to outsmart detection DOI

Taseef Ayub,

Rayees Ahmad Malla,

Mashood Yousuf Khan

и другие.

Global Knowledge Memory and Communication, Год журнала: 2024, Номер unknown

Опубликована: Авг. 14, 2024

Purpose The study aims to investigate the influence of HIX.AI, an artificial intelligence (AI) tool that humanizes generated content, on detection capabilities AI-generated text detectors. Design/methodology/approach investigates reliability six content tools by passing ten essays, five each using Chat Generative Pre-Trained Transformer (ChatGPT) and Bard (Gemini) before after through which content. Findings found selected detectors identified with inconsistencies. Some essays were falsely as human-written a few detectors, indicating are unreliable. Post-HIX.AI application all passed except two, mixed two separate Practical implications findings present evolving field can bypass highlighting difficulties in identifying presence humanization tool. Passing has serious consequences, especially academics. Hence, recommends more robust distinguish accurately. Originality/value contributes existing literature AI highlights challenges pose

Язык: Английский

Towards an AI policy framework in scholarly publishing DOI
Zhicheng Lin

Trends in Cognitive Sciences, Год журнала: 2024, Номер 28(2), С. 85 - 88

Опубликована: Янв. 8, 2024

Язык: Английский

Процитировано

10

Towards an AI policy framework in scholarly publishing DOI Open Access
Zhicheng Lin

Опубликована: Окт. 1, 2023

Generative artificial intelligence (AI) tools are rapidly transforming scholarly research. To harness their benefits while minimizing potential pitfalls, robust policies urgently needed. In analyzing AI for authors and reviewers from major publishing organizations leading journal groups, we found not only marked inconsistencies in permitted usage reporting requirements, but also a general lack of clarity or guidance on ethical principles, transparency, reproducibility. Furthermore, the restrictions set by these largely unenforceable. establish norms best practices, propose pragmatic, enabling principle that encourages to produce work with assistance, emphasizing responsible disclosure over unenforceable restrictions. Building this strengths existing policies, offer ready-to-use author reviewer policy templates. When systems used beyond typical editing services, should report details tool, its usage, generated content, prompts. Reviewer articulate confidentiality risks. Our analysis templates roadmap building consensus practices integration into routines, thereby accelerating discovery without compromising academic integrity.

Язык: Английский

Процитировано

4

The art of deception: humanizing AI to outsmart detection DOI

Taseef Ayub,

Rayees Ahmad Malla,

Mashood Yousuf Khan

и другие.

Global Knowledge Memory and Communication, Год журнала: 2024, Номер unknown

Опубликована: Авг. 14, 2024

Purpose The study aims to investigate the influence of HIX.AI, an artificial intelligence (AI) tool that humanizes generated content, on detection capabilities AI-generated text detectors. Design/methodology/approach investigates reliability six content tools by passing ten essays, five each using Chat Generative Pre-Trained Transformer (ChatGPT) and Bard (Gemini) before after through which content. Findings found selected detectors identified with inconsistencies. Some essays were falsely as human-written a few detectors, indicating are unreliable. Post-HIX.AI application all passed except two, mixed two separate Practical implications findings present evolving field can bypass highlighting difficulties in identifying presence humanization tool. Passing has serious consequences, especially academics. Hence, recommends more robust distinguish accurately. Originality/value contributes existing literature AI highlights challenges pose

Язык: Английский

Процитировано

0