Maybe you will say justified true belief, but some claim that it doesn't account for Gettier's, so: is there a better definition which is becoming more prominent today? Or can the JTB, perhaps with some slight modification, overcome the problems posed by Gettier?
One thing seems plain to me: knowledge is a subset of belief and it must be justified. I've thought that maybe knowledge does not require truth, but then it is something of a [approximately] justified [approximate] certainty... it feels a bit awkward. However if it does necessitate truth, then hardly can we ever possess knowledge.