Wasim Wahid

I would suggest reference [1] as introductory reading .

Now, to address the question with my own view:

I do not see any further elucidation of adding AI into the ethical evaluation. The reason being is that current AI systems have no independence from its creators. Therefore the creators are all responsible ethically for the AI (even those AI engines such as ChatGPT). I hold this view based on the fact that these are not second order systems. That is, they cannot discriminate the information that is being fed into them. They just train on the data given and output what was on the data. The data was selected by the designers and therefore it is the designers responsibility. Any argument that they are black boxes is a weak effort to try to evade responsibility for a flawed design. Programmers and data scientist should be held responsible for their systems if they do not take reasonable precautions and hold off on implementing systems they do not understand.

References

[1]Moral agency without responsibility? Analysis of three ethical models of human-computer interaction in times of artificial intelligence (AI) by

Alexis Fritz, Wiebke Brandt, Henner Gimpel and Sarah Bayer

Dear Wasim Wahid ,

AI systems can perpetuate existing biases and discrimination in society. The algorithms are only as unbiased as the data they are trained on. Thus, it is necessary to ensure that the data is diverse and representative of different groups to prevent discrimination.

https://www.unesco.org/en/artificial-intelligence/recommendation-ethics/cases

Regards,

Shafagat

Most of the research or studies are normally focused on the development of strategies for decision-making using AI. Though the model's accuracies are normally tested but if the data itself is biased or contain discriminatory patterns, the AI models can perpetuate or amplify those biases. This can lead to unfair outcomes and discriminatory decision-making. Now there is a big need to divert the research toward the evaluation of the decision-making, The PDCA cycle can best explain the process. For example, if the decision-making mechanism has been established and implemented, then there is a need to evaluate the decision-making process by analyzing the dataset of maintenance or management. So there is a need to focus on this part as well.