I think it's to allow for the writing of equations that then are subsequently used in Solve, DSolve, and other such functions. If Equal were to return False when both sides do not explicitly evaluate to True or False then there'd be no mechanism with which to write systems of equations. For example the following would not be possible as a useful syntax

Solve[a x + b ==c, x]

As a general rule I always use SameQ if I want to make a determination of truth in an expression. However, you might want to use Equal to allow for the possibility that an expression actually might not evaluate to either True or False and thereby make use of (for example) a third argument in an If statement.

When I look at other people's code I find they often do not make use of === and never notice the potential issue of using == until it bites them. An alternative is to make use of == along with wrapping the expression in TrueQ which forces a False when the expression using == does not evaluate to True or False. So--rule of thumb--always use === over == when doing a test of truth unless there's a good reason to make use of ==.

Thanks for the quick and simple answer, I am still confused on what that means for how the language is written.

As I understand it:

SameQ is a strict equality test that returns false unless it returns true. Equal is a looser equality test (allows for type differences) but doesn't automatically return false.

Do you know why Mathematica does not return a false in this when == is used? What advantages does this give?

Is this because Equals will sometimes try to return true when they are slightly different, and want to avoid returning false negatives?