Fallibilism : evidence and knowledge

What strength of evidence is required for knowledge? Ordinarily, we often claim to know something on the basis of evidence which doesn't guarantee its truth. For instance, one might claim to know that one sees a crow on the basis of visual experience even though having that experience does not guarantee that there is a crow (it might be a rook, or one might be dreaming). As a result, those wanting to avoid philosophical scepticism have standardly embraced 'fallibilism': one can know a proposition on the basis of evidence that supports it even if the evidence doesn't guarantee its truth. Despite this, there's been a persistent temptation to endorse 'infallibilism', according to which knowledge requires evidence that guarantees truth. For doesn't it sound contradictory to simultaneously claim to know and admit the possibility of error?