I suspect the claims were distorted versions of what was found in the study below. It supports a weaker form of the above claims: the idea that a person would often look at someone they are socially close to when laughing in a group setting, but not necessarily at the socially closest person.
From McAdams, et. al. (1984):
[High scores on] intimacy motivation was positively associated with
greater levels of laughter, smiling, and eye contact.
Along with non verbal behaviors [...], looking, laughter, and smiling have been classified as cues of [...] "intimacy".
Harper et al. (1978) conclude: "With some exceptions and qualifications, the research to date shows a direct relationship between looking and positive interpersonal sentiment" (p. 189)
Rubin (1970) found that the amount of time in which the couple engaged in mutual eye contact during a laboratory session was positively correlated with their score on a questionnaire measure of love intensity.
Observing adults interacting in dyads, Rosenfeld (1966) found that the frequency of smiling significantly differentiated between those subjects instructed to seek approval and those instructed to avoid approval from the other.
Anecdotally, it would seem very strange that in a group setting with multiple acquaintances (for example at a dinner table with friends/family), two people would only look at each other every time a joke is made. In settings where social closeness disparity is larger (e.g. at a comedy club show) it seems obvious that two people who went to the show together would look at each other when a joke is made.
The larger point is the usefulness of the laughter-eye-contact response. The response would be useful to gauge more general "who knows who" information, but would probably not be specific to any previous or desired sexual encounters.