This article presents three privacy dark patterns observed in identity management. Dark patterns are software design patterns that intentionally violate requirements, in the given case privacy requirements for identity management. First, the theoretical background is presented, and then next, the observed patterns are documented, described and formalized. The resulting dark patterns show how security is used as obfuscation of data collection, how the seemingly harmless collection of additional data is advertised to end users, and how the use of anonymization technology is actively discouraged by service providers.