Page 413 - Ai_V3.0_c11_flipbook
P. 413

European Union’s Ethics Guidelines for Trustworthy AI
                 Focus: Ethical guidelines for AI development and deployment in the EU
                 Components:

                    • Rules for AI you can trust, like letting people make their own choices, avoiding harm, being fair, and taking responsibility.
                    • Needs for AI to be clear, understandable, and able to be checked in detail.
                    • Ideas for making sure people are incharge and there are ways to check and take responsibility for AI that affects
                   society a lot.



                               Brainy Fact


                      IBM’s AI Fairness 360 is an open-source toolkit designed to address bias in machine learning models. It
                      includes over 70 fairness metrics to help users detect bias, indicating its robust capability for identifying
                      potential bias sources. The toolkit also offers more than 10 algorithms for mitigating bias, such as
                      optimising the preprocessing stage, prejudice remover, and regular algorithms. With its comprehensive
                      features, educational resources, and validation mechanisms, AI Fairness 360 aims to promote fairness
                      and equity in AI applications.
                      Building Trusted AI pipelines

                      (Using Open-Source)
                      Was it tampered with?         Is it fair?      Is it easy to understand?  Is it accountable?









                          ROBUSTNESS                FAIRNESS            EXPLAINABILITY             LINEAGE



                        AI and Ethical Concerns


                 Ethics is defined as the discipline that deals with human moral obligations and duties. It is a set of ethical principles that
                 regulate the behaviour and actions of individuals or groups. Artificial Intelligence ethics is a part of ethical technology
                 specific to robots and other artificially intelligent machines. Several ethical concerns regarding AI include:

                    • If AI produces human-like results, can it also take human-like decisions?
                    • If AI decides whether to grant bank loans, will the AI algorithms be fair and just?
                    • If AI makes human decisions, is it also worthy of human confidence?
                    • AI is basically data + mathematical model + data-based training + prediction. What if the data provided for training
                   is unintentionally skewed or biased?
                    • Will AI lead to significant job displacement?
                 These questions and concerns come up all the time, so the “ethics of Artificial Intelligence” is very important. We must
                 understand how Artificial Intelligence works and the framework of Artificial Intelligence ethics.


                        Ethical Dilemma

                 An ethical dilemma is a situation in which a person faces a choice between conflicting moral principles or values. It often involves
                 tough decisions where there are competing interests or where doing what is considered right may result in undesirable outcomes.

                                                                                         AI Ethics and Values   411
   408   409   410   411   412   413   414   415   416   417   418