In writing computer programs, do too many cooks spoil the broth ?

Suppose, for example, model-based testing is being done. The more people on the team, the more likely at least one person on the team will misconstrue what is being done. Instead of ensuring the program agrees with the model, he will begin doctoring the model to reflect the released code. Without some sort of anti-counterfeiting technology to show who is doctoring the code, the person working correctly is simply aiding and abetting the fool, who will go on to gain promotions and awards for the defective and unsafe system that is inevitably being built.