Since around July 2007, Oberwiki has used a CAPTCHA test to verify that edits and new accounts are made by people, and not by automated spamming bots. (In the past, a combination of word blacklists and IP address banning was used.)
A CAPTCHA test is a text-recognition test that is relatively easy for a human to pass, but relatively difficult for a computer to pass. Traditionally, the test consists of an image of distorted text that the user must correctly type.
The particular implementation that Oberwiki uses is called reCAPTCHA, and was developed by Carnegie Mellon U. It uses two words, both of which come from a book digitizing project, such as the Internet Archive's Text Archive. One word is known, but the other is difficult for a computer to read. When a user types the text of both words, it gives the project a suggestion as to what the unknown word is, thus helping the digitizing project.
Though it doesn't sound like much help, the total time spent solving CAPTCHAs works out to around 150 thousand human hours daily.
For users who cannot read the words, there is an audio test.