According to Google cloaking is "a website that returns altered webpages to search engines crawling the site." In other words, a human reading the site would see different content or information than the Googlebot or other search engine robot reading the site. Most of the time, cloaking is implemented in order to improve search engine ranking by misleading the search engine robot into thinking the content on the page is different than it really is.
Most search engines will immediately remove and sometimes blacklist a site that is discovered to be cloaking. They do this because cloaking is usually intended to completely fool the search engine's algorithms and programming that determine what makes a site rank high or low in that engine. If the page that the customer sees is different from the page that the search engine bot sees, then the search engine cannot do its job. So they ban sites that use cloaking.
What's At Stake?
Is Personalization Cloaking?
One of the newest features of many advanced Web sites is to display specialized content depending on various factors determined by the customers themselves. For example, on About, if you haven't visited the site in several months, you might get different content displayed in the navigation menus than you would if you regularly visited the site. Other sites use a technique called "Geo-IP" which determines your location based on the IP address you are logged into and displays ads or weather information relevant to your part of the world or country.
Some people have argued that this personalization is a form of cloaking, because the content that is delivered to a customer is different than what is delivered to the search engine robot. But the robot receives the same type of content as the customer, just personalized (if you will) to that robot's locale or profile on the system.
If the content you are delivering is not dependant upon knowing if the visitor is a search engine robot or not, then the content has not been cloaked.