A series of experiments conducted in Japan by Yamagishi and his associates are presented, all consistently showing that high trusters (as measured with a general trust scale) are more sensitive than low trusters to information potentially revealing lack of trustworthiness in others and judge other people's choice in a one‐shot prisoner's dilemma more accurately. Based on these findings, a new theoretical twist is introduced to the “emancipation theory of trust” originally proposed by Yamagishi & Yamagishi (1994), that emphasizes the relation‐expansion role of trust in addition to the traditionally noticed relation‐fortification role of trust. When opportunity cost for staying in a commitment relation is generally high, it is more advantageous not to stay in secure and stable commitment relations but to explore opportunities that lie outside, and yet such social exploration involves the risk of being exploited by untrustworthy people. It is thus a more gainful strategy to invest “cognitive resources” in the nurturing of “social intelligence” needed to detect signals of untrustworthiness. General trust may be conceived as a by‐product of the development of such social intelligence. Those who have invested in the development of social intelligence can afford to maintain a high level of general trust, whereas those who have not are encouraged to assume that “everyone is a thief” and to refrain from pursuing potentially lucrative but risky outside opportunities.