This form of AI deals with an entity in an environment where the entity know or gathers certain information about what is around it. This information, however, might be false information or half-truths. Think of optical illusions. They trick us into believing certain things through sight, but later we find out the truth behind the illusion. Not everything we see is a trick, so we usually believe what we see. From this information we form desires. Once we have desires (we threw out all the bad ones or ones we don’t want to pursue) we form intentions on how to obtain our desires. This is the basis for the BDI architecture. Suppose we think we see a beautiful woman (or man if you are a woman trying to relate to this example) from a distance. We believe this to be an attractive person. We desire to confront the person and form a relationship. We intend on talking to the person and asking them for their phone number in order to obtain these desires. This architecture, like many other architectures, have the ability to adapt when new information is received. Suppose that once we get closer to the person we saw earlier from a distance, we realize that the person is not quite as attractive as we thought they were (time to go to the opomologist). Our beliefs change, which then changes our desires and intentions.