A Survey of Experts on General Intelligence, Race Differences, and IQ Testing

From the abstract:

Results indicate expert consensus that g is an important, non-trivial determinant (or at least predictor) of important real world outcomes for which there is no substitute, and that tests of g are valid and generally free from racial bias.

 and,

Experts did not reach consensus on issues such as the degree to which specific abilities or combinations of non-cognitive traits can yield predictive validities comparable to that of g alone, the predictive validity of g for non-technical work outcomes (e.g., contextual performance), and the nature and implications of race differences in intelligence.