Se copulas were not recorded at all in Santo/Malo in Meyerhoff's (1994–1995) fieldwork. I heard it several times in Vila in 1998, 2003. It may be regionalized or may have originated in Vila. I am not at all sure how widely se is accepted as predicate copula, so more research on this emergent construction would be valuable.
Crowley (1989) suggests the se copula originated by analogy with French c’est. If this is true, and if the functions of French c'est have influenced its grammaticalization in Bislama, se may first emerge in Bislama in predications that seem somehow 'topicalized'. If it comes from c'est, it may also be preferred by speakers with French education, but this is unknown. It may be preferred as an identificational copula but this is also uncertain.
Source: nd
Source: Meyerhoff nd