“Reports a specific research finding on human-AI interaction with objective language, avoiding value judgments on dependency.”