Anzeige
Erschienen in:
2023 | OriginalPaper | Buchkapitel
3. Ethics and Moral
verfasst von : Josef Baker-Brunnbauer
Erschienen in: Trustworthy Artificial Intelligence Implementation
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Abstract
This chapter focuses on ethics and moral based on an AI environment. It sums up that all past societies had ethical standards with the central goal to survive. Ethics itself goes back a long time in history. Already Aristotle was speaking about ethics, and it will be an even more central topic for the future of humanity. Will AGI consider the needs and understandings of the lower-developed humans? How will humans handle “intelligence explosion” and recursive scenarios, where AI or AGI will replicate itself constantly? Humans reshaped the planet earth to gain benefits, will this be done by an AI system as well?
This chapter lists an overview of different AI ethics frameworks and lists different viewpoints for the situation that AI systems will have their own moral status. Section 3.1 describes approaches to how to train AI systems in ethics. One major challenge is, if AI systems will get trained with already biased data sets. Overall, it is still unclear how to teach ethics to AI systems. Section 3.2 lists criteria for Product Development of AI systems and describes the importance of testing environments. Next, Sect. 3.3 analyses some ethical and moral challenges. It shows that moral decision-making is not about logic and rationality, instead, it is about psychologically acceptable explanations. Therefore, it requires a deep understanding of humans and their behavior.
Other challenges are accountability (who is responsible for a failure?), law adaption, quick reproduction of AI systems, and the not predictable social impact. Section 3.4 dives deeper into existing guidelines. A survey shows the importance of the requirement of transparency and safety for users and consumers. It describes the outcome of the summary of an overview of 47 values and principles that are based on several manifestos. The AIHLEG of the European Commission defines Trustworthy AI by seven key requirements and offers an AI assessment list.