Deceiving ourselves into believing our own lies is probably evolutionarily beneficial. After all, everyone does do this on some level. According to Robert Trivers, an evolutionary biologist, we believe our own lies in order to better deceive others. This way we can excel in various social situations with very little effort. This makes good sense, until you take into account the lack of falsifiability that is present in all evolutionary hypotheses.
Then again, there is evidence to support the idea that self deception has been beneficial to our survival. Sackeim and Gur (1979) devised categories or deception to test the amount of impact that self deception has on people. They found that self deception held more influence over most people than any other form of deception that they measured. That this was true for their participants suggests that it is true for humans in general, and this would suggest that there was some evolutionary advantage to the behaviour (and perhaps this advantage is still present). However, this study only had 250 participants so can’t possibly be generalised to wider society, let alone across the history of our species. Another problem with this experiment is that the categories that they created may not be ecologically valid.
Trivers has also said that self deception has the drawback of making people more susceptible to manipulation by others. The idea that humans are intrinsically flawed is supported by Forgas and Williams (2003), who say that our judgement among other things is terribly imperfect. This doesn’t go against the evolutionary approach, as evolution doesn’t create perfect species, and the fact that most people have the tendency to deceive themselves shows that evolution has at the least kept this trait alive by coincidence. I disagree that there is necessarily an advantage to every trait that human beings have, though, which is what the evolutionary approach seems to put forward.