“Creating” an audit to check the ethics of Artificial Intelligence

Abstract

In this paper we focus on creating an audit to verify the ethics in artificial intelligence (AI) as it turns out to be the third pillar of AI alongside legal and technical issues. Precisely because of the importance of auditing the ethics behind AI decisions, the author uses ChatGPT to propose an ethical audit based initially on addressing eight essential questions, namely: How is it ensured that the data used to train the AI is free from discriminatory biases?; What mechanisms are in place for users affected by AI decisions to appeal biased decisions?; What measures are implemented to prevent models from generating offensive or harmful content?; How is the use of AI managed to prevent its abuse in creating deepfakes or manipulating images and videos?; How is it ensured that the data used to train the AI complies with data protection regulations such as GDPR?; How is the environmental impact of training AI models managed?; What measures are implemented to ensure that AI is not used for malicious purposes?; What measures are taken to protect minors from exposure to AI that may negatively influence their development? The goal of the paper is to lay the foundations for an AI ethics audit, ensuring that ethics is integrated into the DNA of any designed AI.