Australian research has shown how easily eye witness accounts, a foundation stone of the judicial system, can be "contaminated".
The University of Sydney study found people could add new detail to, or even amend, their memory if they discussed their recollection with someone who saw the same event.
It was a subconscious process, said Dr Helen Paterson, whose research has focused on chatter between co-witnesses and its role as a "potent delivery mechanism" for false memories.
News that makes sense
Your trusted source for staying up-to-date with the world around you. Get free daily news updates and analysis, straight to your inbox.
"A false memory is the recollection of an event, or details of an event, that did not actually occur," Dr Paterson said on Monday.
"Witnesses who discuss an event with a co-witness are very likely to incorporate misinformation presented by the co-witness into their own memory for the event."
Dr Paterson said when two people discussed an event they had both witnessed they could then find it "difficult, if not impossible" to sort their genuine memories from information they were told.
When deliberate misinformation was introduced, even warning a witness about this made them no more able to sort the fact from the fiction.
Dr Paterson also said discussions between co-witnesses were found to have more affect on a person's memory than exposure to inaccurate media reports or leading questioning.
And once a false memory was implanted, it could be very difficult to shake.
"Once their memory has been contaminated in this way, the witness is often unable to distinguish between accurate and inaccurate memories," Dr Paterson said.
Dr Paterson's research uses a video of a house burglary that test subjects watch on their own without knowing there are two different versions.
After a co-witness discussion, most participants go on to report remembering specific details that were not in the version they were shown.
She said the research underscored the need for police and the courts to discourage contact between co-witnesses who were usually barred from hearing each others' testimony in court.
"Despite these attempts to keep witness testimony independent, it is clear that witnesses often do talk to each other about the event," Dr Paterson said.
"Discussion among witnesses is difficult, if not impossible, to prevent."
