Life expectancy from birth is easily the most misleading statistic in the history of the social sciences because it is a measure of central tendency (aka an average, specifically, a median) of a property (age at death) that not only has no central tendency but actually has the opposite of a central tendency, with values concentrated at the low end (infant and child mortality) and the high end (old age deaths). In almost all societies ever measured, the life expectancy from birth age is usually the age at which a person is least likely to die.

To add to its misleading nature: demographers usually use the value to express the life chances of the just-born cohort (up to age 5). Since they obviously can't wait 70 or 80 years until half of that cohort has actually died, they instead use curve-fitting to estimate life expectancy based on infant and child mortality actually experienced by the cohort. People often say that life expectancy from birth is misleading because it's heavily impacted by infant and child mortality, but this is not quite correct - it's actually entirely determined by infant and child mortality.