Technology ethics: data, algorithms, and responsibility
New technologies create new choices, and with them new questions about privacy, fairness, and who is to blame when things go wrong.
Privacy and data
Apps and devices collect huge amounts of personal information, often as a condition of use. Ethical questions include whether clicking 'agree' on a long policy counts as meaningful consent, who owns data about you, and whether information that is harmless alone becomes dangerous once combined. Privacy protects people's ability to control how they appear to others.
Algorithmic bias
Algorithms trained on historical data can repeat past unfairness. A hiring tool trained on a company's previous hires may learn to favour applicants who resemble them, even if no one programmed it to discriminate. Because such systems are fast and appear objective, their errors can spread widely and be hard to challenge.
Responsibility gaps
When an automated system causes harm, responsibility can seem to vanish: the designer did not intend it, the user did not control it, and the machine is not a moral agent. Philosophers call this a responsibility gap. Proposed solutions include clear accountability rules for developers and deployers, human oversight of important decisions, and transparency about how systems work.
Flashcard
Systematic unfairness produced by an automated system
Check your understanding
A loan algorithm trained on past approvals keeps rejecting applicants from one neighbourhood that banks once unfairly avoided. This is best described as: