Brewin et al. conclude that memory is generally reliable unless influenced by external factors.
Chris R. Brewin, Bernice Andrews, and Laura Mickes, "Regaining Consensus on the Reliability of Memory," Current Directions in Psychological Science 29, no. 2 (2020)
Conclusions
The claims examined here, repeatedly cited in textbooks and articles, paint a picture of memory as being highly prone to error and suggestion to an extent that the public is unaware of. All three, however, rest on research with methodological weaknesses that invite alternative interpretations of the findings. Reanalysis of this research and the collection of new data suggest that more modest conclusions are appropriate. One implication is that research on the “dark side” of memory has not yet provided grounds to overturn the long-held consensus of memory as essentially doing a good job, despite being limited by its capacity and prone to manipulation. Memory is clearly malleable but not unreliable under normal circumstances and in the absence of contamination or prolonged suggestion by psychologists, therapists, or anybody else (Koriat et al., 2000; Wixted, Mickes, & Fisher, 2018).
Another implication is that psychologists should be much more cautious in assuming that their understanding of memory is superior to that of the ordinary person and that institutions such as the criminal justice system need educating about research. This may well be true in some areas, but it fails to consider the possible costs of disseminating findings that do not rest on a solid foundation. Ironically, when questioned about the relation between accuracy and confidence, nonexperts have been found to give answers that are more consistent with the latest research than the ones given by psychology students, who, rather than probing their own experiences, had been misled by the conclusions in the literature (Brewin et al., 2019).
Ultimately, psychologists are responsible for the way in which our research is interpreted by the outside world. There is a real danger that overenthusiastic championing of conclusions based on limited data, without proper regard for methodological caveats or issues of external validity, will be damaging both to our reputation as scientists and to potential recipients, such as legal professionals, who rely on us to guide them with restraint and impartiality.