The ethics and governance of Artifi cial Intelligence large language models

Zhihong Liu

doi:10.59429/esta.v10i4.1601

Zhihong Liu

DOI: https://doi.org/10.59429/esta.v10i4.1601

Keywords: artificial intelligence, large language models, ethical security, governance.

Abstract

After the construction of large language artifi cial intelligence (AI) models and their provision as a social service, potential risks quickly become real risks that aff ect our society. In the past, we have always responded to ethical and safety issues with AI in a reactive manner, rather than proactively. As a result, we encounter some of the social problems brought about by large language AI models. This paper aims to focus on the existing problems caused by large language AI models and promote the implementation of ethical and safety governance.

References

[1] Zero-experience - Four hours! Researchers Use ChatGPT to Build a VirusTotal Zero-Detection Steganography Malware� EB/OL�.http://www.hackdig.

com/04/hack-958424.htm, 2023-04-06

[2] https://www.darkreading.com/attacks-breaches/researcher-tricks-chatgpt-undetectable-steganography-malware

[3] https://www.forcepoint.com/blog/x-labs/zero-day-exfi ltration-using-chatgpt-prompts

[4] Nathan George. Practical Data Science with Python: Learn tools and techniques from hands-on examples to extract insights from data[M]. Packt

Publishing Private Limited,2021

[5] Emergency! Trillion giant moves: Forbidden! [EB/OL]. https://new.qq.com/rain/a/20230502A05MGI00, 2023-05-02

[6] https://www.cs.purdue.edu/homes/ninghui/papers/ t_closeness_icde07.pdf