It’s not clear whether or not her comments are based on research, but the claim seems to be that web 2.0 brought about a shift whereby people went online ‘to be with people they already know’. This reinforces divisions along ethnic lines.
The argument is backed up with recognisable examples of people from ethnic groups explaining that their online connections reflect their own culture and background. Hispanic and African-American people use Twitter disproportionately; white women use Pinterest; Asian Americans use Tumblr. Partly, as one participant in the video points out, this could be explained by clusters of early adopters occupying a space, and the homophily principle (birds of a feather) applies.
So what would we expect? Lee Rainie and Barry Wellman in their recent book make the point that social media
‘enhances the density of inter-connections among a person’s relatively close ties by allowing friends of friends to become aware of each other.’
The argument is surely that social media helps strengthen people’s weaker ties far more than was possible for most of us pre-internet. Weaker doesn’t necessarily mean ‘weak’. So it’s probably a question of expectations. If you were led to expect that that would mean demonstrably increased diversity, and that social media could overcome homophily, you may be disappointed.
And as this recent Atlantic Cities article points out, a linear model of diversity and segregation is too simplistic: ‘you can have segregation and diversity in the same place, at the same time.’