Whilst
Ophiocordyceps unilateralis has been colloquially named as creating zombie ants if we take a few minutes to read some peer-reviewed papers we see that this name doesn't reflect the reality. Firstly there are many examples of parasites that change behaviour;
- The http://en.wikipedia.org/wiki/File:Dicrocoelium_LifeCycle.gif" involves it infecting ants and making them climb to the tallest blades of grass so that they can be eaten by cattle.
- Spinochordodes tellinii must reproduce in water but grows in insects such as grasshoppers. Once fully grown the parasite will encourage the grasshopper to jump far into the water so that it can be released and reproduce (interestingly it doesn't cause the grasshopper to seek water, rather when it sees it it jumps
- When a rodent consumes droppings containing Toxoplasma gondii it has it's fight or flight response disrupted so a cat can just walk up to it and eat it. This allows the parasite to http://en.wikipedia.org/wiki/File:Toxoplasmosis_life_cycle_en.svg" .
As for the so called Zombie ant fungus its behavioural changes are nothing new. The fungus begins to replicate and causes a shift in behaviour resulting in the ant leaving the colony to find a suitable spot. It will climb onto a leaf or stem and trigger an intense clamping action with it's mandibles (there's a picture in one of the videos you linked of an ant doing this to another ant but this is an unusual accident). Then the ant dies allowing the fungus to grow and sporulate, however the spores of
O. unilateralis cannot be carried by the wind and so just settle in the surrounding area. Other ants that walk through then become infected and perform the same death activity nearby. The result is areas called "graveyards" where for tens of metres plants are covered in dead fungus ants.
For more info http://www.plosone.org/article/info:doi/10.1371/journal.pone.0017024" of infected ants.
As interesting and disgusting as this is it is a far cry from reanimated corpses running around to feast on the living I'm afraid. Also parasites evolve in very intense and complicated life cycles with multiple hosts at multiple life stages. It's therefore very science fiction to suggest that some clandestine government agency would be able to engineer a virus to work in humans; when they've evolved for the biology and nervous system of several different animals many of which not mammals adapting it to humans for zombification purposes would pretty much require you to invent such a disease from scratch.
Having said that the rabies virus can induce rage like behaviour with diminished cognitive function but it also causes flu like symptoms, delirium and death.