AI in High-Stakes Decisions: Medicine, Law and Money
Some decisions are ordinary: which song to play next, which advert to show you. If the computer gets those wrong, you shrug and move on. But other decisions can change or even end a life: Does this patient have cancer? Should this person stay in jail? Can this family get a loan or keep their benefits? These are high-stakes decisions, and AI systems are increasingly used to make or shape them.
This is one of the most important and most controversial areas in all of AI ethics. The honest answer is not "AI is good here" or "AI is bad here." The truth is a genuine tension: a well-built AI can be fairer and more consistent than tired, distracted, biased humans — and a badly-built one can fail thousands of people at once, silently, in ways nobody notices until the damage is done. Your job in this lesson is to learn how to tell the difference.
What you will learn
- The three tests that make a decision "high-stakes" — and why they raise the ethical bar.
- Real evidence for both sides: where medical, legal and financial AI genuinely helps, and where it has genuinely harmed real people.
- Four core concepts every citizen should understand: false positives vs false negatives, automation bias, opacity, and accountability chains.
- How a proxy can quietly poison a system — walked through with real numbers.
- A design checklist for responsible high-stakes AI, and a framework for deciding when AI assistance is appropriate versus reckless.
- What all of this means for Kyrgyzstan and Central Asia specifically.