What is Ethics in AI?

AI ethics moved from academic philosophy to practical engineering in the past decade as deployed AI systems began causing measurable harm. The major issue areas are well-defined. Algorithmic bias: models trained on biased data produce biased decisions, with documented harm in hiring, lending, healthcare, and criminal justice. Privacy: AI systems often require massive personal data, creating surveillance risks and breach exposure. Transparency and explainability: black-box models making consequential decisions undermine due process and informed consent. Accountability: when AI systems cause harm, who is responsible — the developer, the deployer, the user, or no one? Labor displacement: AI is reshaping work in ways that distribute benefits unequally. Misinformation and synthetic media: deepfakes and AI-generated content erode trust in evidence. Concentration of power: a small number of companies control frontier AI systems with global influence. AI safety: research community concerns about advanced AI systems pursuing unintended goals. Frameworks for addressing these include regulatory approaches (EU AI Act, US executive orders), industry standards (NIST AI Risk Management Framework, ISO 42001), and internal practices at AI labs (constitutional AI at Anthropic, OpenAI's preparedness framework). Ethics in AI is not about whether to build AI — it's about building it in ways that distribute benefits broadly, surface and mitigate harms, and preserve human autonomy and oversight at consequential decision points.