If unambiguous samples are found and properly handled, it can be.
Here is what often goes wrong:
1) A person is buried 900 years ago. 400 years later, little is left of the remains. A tree root invades the burial site, and in time, the tree dies, and the root decays. A sample of material surrounding the bones is counted.
What is the indicated age? Some average of the two.
2) An unfired clay figurine had been reinforced with sticks during its making. So it is safe to assume that the sticks date from around the time the clay was formed. However, the clay was taken from a riverbed, where, a thousand years before, plant matter was entrapped in the clay, laid down in a flood season. The result is artifically older. C14 has a half life of 5730 years +/- 40 years and this rate is absolute and definite, but sample mixing can interfere badly.
In Egyptian tomb cases, where food offerings have been buried in pottery jars, it is less likely that thousand-year old grain was buried with the dead, so these dates are confident. But suppose something belonging to an ancient ancestor were buried and measured!
In my part of the country, remains do not last long in Northern Woodlands soil, so the potential for error is high.