"Dangerous" to what? I don't recall a specific relevant discussion, was it about danger to customer lock-in moats or about security or about something else ?
However, I don't think it's a fair comparison. In the CA case, users were deceived into a research and had their social graph being accessed without being told such implication. Here, the user has the control over the generation of links and the resource being linked; the user must manually generate and invoke the link to cross-access resource between apps.
[1] (https://en.wikipedia.org/wiki/Facebook%E2%80%93Cambridge_Ana...)