AI Influence on Human Thought and Action: A Lexicon and Taxonomy of a Fragmented Literature

Keeping a human “in the loop” is supposed to make artificial intelligence safer. AI offers an answer. A person decides what to do with it. But AI can influence the person before that decision is made. It can affect what reaches the person. It can direct attention. It can shape how information is understood. Researchers in several fields have already identified parts of this problem. The findings are scattered under different names. Some terms come from older automation research. Others describe effects made newly important by contemporary AI. I organize these terms into a cross-disciplinary map of AI effects on the human. The map distinguishes what AI tells a person from how AI shapes the setting in which the person thinks. It also places one-time effects alongside feedback loops that can change both human and machine. The central claim is not that reciprocal human–AI influence is newly discovered. It is that a fragmented vocabulary makes related forms of influence harder to recognize and compare. A usable map can clarify what kind of influence is occurring and what human-in-the-loop oversight can—and cannot—guarantee.