Knowra AI ethics AI ethics AI ethics studies moral questions raised by the design, use, and effects of artificial intelligence, including responsibility, fairness, privacy, and harm.
Algorithmic bias : Systematic and unfair differences in algorithmic outcomes across people or groups. Bias raises questions about fairness in data, model design, and automated decisions.
Facial recognition system : A system that identifies or verifies people by analyzing facial images. Its deployment in public spaces brings privacy, accuracy, and surveillance concerns together.
Algorithmic discrimination : Unjust treatment of people produced or reinforced by algorithmic systems. It describes a direct social harm that fairness assessments seek to detect and prevent.
AI safety : Research and practice aimed at preventing AI systems from causing unintended or unacceptable harm. Safety emphasizes risk reduction, while AI ethics also examines justice, rights, and power.
Explainable artificial intelligence : Methods that make an AI system’s reasoning or outputs more understandable to people. Explanations can support scrutiny, accountability, and informed challenges to AI decisions.
Predictive policing : The use of data analysis to forecast where crime may occur or who may be involved. Historical enforcement data can reproduce unequal policing patterns in new predictions.
Automation bias : The tendency to trust automated recommendations more than the evidence warrants. Human oversight can fail when people defer uncritically to AI outputs.
Responsible AI : A set of practices for developing and deploying AI with attention to social and ethical impacts. It turns ethical principles into organizational processes and technical controls.
Privacy : The ability to control access to personal information and aspects of one’s life. AI systems can infer sensitive facts or expose information through data collection and use.
Automated hiring : The use of software to screen, rank, or select job applicants. Hiring tools can affect access to work while hiding how candidates are assessed.
Show all 23