Mollom

Mollom automatically blocks comment form spam, contact form spam and fake user accounts using a filtering technique based on the combination of content analysis and CAPTCHA challenges. When new content is analyzed by Mollom's intelligent text-analysis filter, and Mollom is unsure whether it is ham or spam, it asks the user to answer a CAPTCHA challenge. This challenge-response procedure doesn't block human users. If an unwanted message still makes it onto a website, users can help fight back by reporting to Mollom. The service learns from its mistakes.