Interesting question and one I have thought a lot about. I hold a moral anti-realism view and think most people speaking about "morality" are discussing normative views.  For this reason, a Virtue Ethics framework would translate really well into a better world under your scenario. A framework designed to optimize people being better rather than doing better would hopefully be an improvement a step earlier in the process of living. I am not familiar with a lot of criticism about virtue ethics, but open to reconsidering. Additionally, a moral anti-realism position isn't necessary for virtue ethics to be an ideal framework in my mind. 

I'd be a bad god.  I'd probably encode some mix of kindness and responsibility, and likely a more static enjoyment of what is than a striving for change.  I presume they'd never get out of hunter/gatherer mode.

And now I'm wondering exactly what my limits are.  I don't think "a theory of morality" is something that stands alone in people.  I translated that in my mind to a set of personality traits and behaviors that would automatically be enforced somehow and not evolve over time (in individuals or across generations).  But if you mean more constrained cognitive moral theories that most of 'em don't actually follow very well, I'm not sure what I'd choose.

Note that none of this applies to real humans, nor perhaps any agents in this universe.  

As evolved—and evolving—agents, we would benefit from increasing awareness of (1) our values, hierarchical and fine-grained, and (2) our methods for promoting those present but evolving values in the world around us, with perceived consequences feeding back and selected for increasing coherence over increasing context of meaning-making (values) and increasing scope of instrumental effectiveness (methods). Lather, rinse, repeat…

As inherently perspectival agents acting to express our present but evolving nature within the bounds of our presently perceived environment of interaction, we can find moral agreement as if we were (metaphorically) individual leaves on the tips of the growing branches of a tree, and by traversing the (increasingly probable) branches of that tree toward the (most probable) trunk, rooted in what we know as the physics of our world, finding agreement at the level(s) of those branches supporting our values-in-common.

I am not a god, but this is the advice I would provide to the next one I happen to meet, and thereby hope to expedite our current haphazard progress (2.71828 steps forward, 1 step back) in the domain of social decision-making assessed as increasingly "moral", or right in principle.

The arrow of morality points not toward any imagined goal, but rather, outward, with increasing coherence over increasing context.

Foundational questions to ponder: am I really God, or do I just think I'm God? How would I test this premise? I'd take a very long time to figure this out. Do I (or the humans) incur any penalty for a delay in encoding morality?

Also, are the humans in question subject to forces of evolution or are we talking about a static landscape? If we mean literal Homo Sapiens, then whatever we encode applies to a finite window as the creatures we manipulate will eventually evolve into something else.