Ethical Risks and Future Direction in Building Trust for Large Language Models Application under the EU AI Act

Abstract

LLMs are being used in an increasing number of AI applications, raising important ethical considerations for which comprehensive guidelines must be framed. LLMs have equal measures of power and risks in bias, privacy breaches, lack of transparency, and the risk of model collapse, eroding the trustworthiness of LLM application. This research examined the ethical risks and pertinent provisions for building trust in LLM applications under the EU AI ACT, which came into force on August 2, 2024. Thus, by concentrating on the compliance aspects related to LLM applications under the EU AI Act, the research encompassed all the relevant aspects of the integration of ethical values with legal norms. The research also offered practical future directions of LLM application for organisations, emphasising the need to perform ethical audits, engage stakeholders, monitor continuously, and enhance AI literacy. The goal of the research is to ensure that LLM applications are deployed responsibly, transparently, and in a manner that upholds public trust, thus contributing to enhancing the overall trustworthiness of LLM applications.