• AI is becoming more vital in the increasingly interconnected world ranging from self-driving cars to automated customer service agents.
• AI is slowly but surely changing how we live and work.
• As AI is becoming sophisticated, the ethical implications of its use are also becoming increasingly complex.
• Several key issues regarding AI ethics, such as data privacy, algorithmic bias, and socio-economic inequality are the biggest challenge to the moral fabric which underpins civilized society.
• Government agencies and policy makers are leveraging AI-powered tools to analyze complex patterns, forecast future scenarios, and provide more informed recommendations.

AI ethics

• Ethics is a set of moral principles which helps us discern between right and wrong.
• AI ethics is a multidisciplinary field that studies how to optimize AI''s beneficial impact while reducing risks and adverse outcomes.
• AI ethics issues include data responsibility, privacy, fairness, explainability, robustness, transparency, environmental sustainability, inclusion, moral agency, value alignment, accountability, trust, and technology misuse. 

Categories of machine agents:

• Machines can be ethical agents responsible for their actions or autonomous moral agents.
• It is categorized by Moore’s classification in 2006:
• Ethical impact agents: machines with ethical consequences that do not make ethical decisions but pose ethical considerations like altering sport’s dynamics such as robot jockeys.
• Implicit ethical agents: machines with embedded safety or ethical guidelines, which follow set rules without actively deciding what is ethical such as safe autopilot system in planes.
• Explicit ethical agents: It goes beyond set rules, using formal methods to estimate the ethical value of options, like systems that balance financial investments with social responsibility.
• Full ethical agents: capable of making and justifying ethical judgments, including reasonable explanations.
• An adult human is a fully ethical agent, and an advanced AI can have similar understanding of ethics.

Ethical challenges:

Philosophical overview:

• Immanuel Kant’s ethical philosophy emphasizes autonomy, rationality, and the moral duty of individuals.
• However, its application to the use of AI in decision-making within governance could lead to serious concerns.
• If decisions that were once the purview of humans are delegated to algorithms, it could threaten the capacity for moral reasoning.
• Isaac Asimov’s ‘Three Laws of Robotics’ implies that a person or institution using AI could be considered to be abdicating their moral responsibility.
• These laws were designed to govern robotic behaviour for ethical actions but can lead to unexpected and often paradoxical outcomes.
• The attempts to codify ethics into rules for robots or complex AI-driven governmental decision-making, are challenging to convert human moral complexity into algorithmic form.
• The intertwining of Kant’s insistence on rational moral agency with Asimov’s fictional exploration of coded ethics illustrates delegating human functions to artificial entities.

Complexity of decision making:

• AI will be used in governance decisions for example decision-making algorithms are used in some countries to determine the beneficiaries of social sector schemes, but it cannot guarantee that decisions assisted by machines remain ethical and moral.
• For example, the programming of a computer to be ethical is much more difficult than programming a computer to play world-champion chess.
• Chess is a simple domain with well-defined legal moves, but Ethics operates in a complex domain with some ill-defined legal moves.
• Governments can delegate a few rudimentary decisions to the machines, but decisions based on machine predictions could be immoral or unethical.
• Artificial Moral Agent (AMA) refers to systems that are more than excellent computers as systems that actually ‘think’.
• The liability of an unethical decision would fall on the developer of the AI system as it is impractical to punish a machine.

Technological advancement:

• Artificial agents are still far from being able to replace human judgment in complex, unpredictable, or unclear ethical scenarios from a technological standpoint.

Bounded ethicality:

• The ability to make ethical choices is often limited because of internal and external pressures, this is bounded ethicality.
• The human patterns of moral disengagement could translate into machine-bounded ethicality.
• Machines may engage in immoral behaviour if framed in a way that detaches ethical principles, similar to humans.
• Moral disengagement in bounded ethical decision-making, allows people to act against their ethics without guilt through moral justifications.
• Many machine predictions are deployed to assist in decisions where a human decision-maker retains the ultimate decision-making authority.

Principles for Ethical AI:

• Policymakers should translate the core values and principles into action with respect to key areas.
• It includes data governance, environment and ecosystems, gender, education and research, health and social well-being.
• UNESCO produced global standard on AI ethics as “Recommendation on the Ethics of AI” which include principles like:
• Proportionality, Fairness and Do no harm
• Right to Privacy and Data Protection
• Multi-stakeholder approach by adaptive governance & collaboration
• Respect International law & national sovereignty
• Responsibility and Accountability
• Awareness and Information Literacy
• Based UN’s Sustainable Development Goals
• AI’s benefits should be accessible to all

The protection of human rights and dignity is based on the advancement of fundamental principles such as transparency and fairness. Responsible AI is an approach to developing and deploying AI from an ethical and legal point of view. Programming ethics into machines is complex, and the world must proceed cautiously. The importance of human oversight of AI systems must be acknowledged at every step of an advancing AI revolution.