The way D&D has often been played, is adventurer paladins condemning "monsters" who detect as evil to death, on the spot, whenever they meet them in a wilderness or a dungeon.
As Quintessential Paladin II (a 3.5 third-party sourcebook that goes into great deal on playing paladins, and potential pitfalls, puts it:
A paladin can detect evil at will. What exactly does this mean?
Leaving aside the mechanical of the spell-like effect for the moment, the real question is, 'what does evil mean in the campaign world?' The paladin's player should discuss this with the Games Master before play begins. After all, a paladin has presumably been training in the use of detect evil for years and knows what it means when a stranger is Evil. Is true Evil (as defined by the spell) common or rare? If someone is evil, does that mean that the paladin should merely be wary of them, should arrest them or should he attack them immediately? This largely depends on the nature of evil in the campaign.
Problems with detect evil normally arise when the paladin is dealing with ordinary people - detecting that a monster in a cave is evil and then slaying that monster never seems to cause moral qualms.
Which is IMO the point The Giant is making with the Sapphire Guard - the moral dissonance, and hypocrisy, of the whole thing - of paladins slaying everything in the wilderness or dungeon that "pings" as Evil on sight, without qualms, and yet being much more careful about due process, in towns.