The report that all espionage charges against Temple University physicist Xiaoxing Xi have been withdrawn raises the obvious question: How could things have gone so wrong?
According to published accounts, prosecutors and agents saw documents that appeared to be suspicious and concluded – erroneously – that they were schematics for a particular device.
Reading the story raises a concern about cognitive biases at work. The term biases does not refer to a prejudice or dislike, but rather a process in which the brain biases the observer to favor a particular conclusion. One particular type, “confirmation bias,” is common and particularly human: What a person expects to see colors the perception of what is then examined.
Confirmation bias is no stranger to criminal and forensic investigations. It played a part in the misidentification of a suspect in a 2004 terrorist train bombing in Madrid, and its effect has been shown repeatedly in research. Whether it is a DNA analyst or a fingerprint examiner, erroneous information or our expectations can cause the person to see what is not actually there or miss what is present.
The concern about cognitive biases is neither illusory nor academic. It is now a subject of discussion at the National Commission on Forensic Science, and forensic labs and police agencies around the country are offering training on cognitive bias and potential systems responses to reduce its effect.
There is no easy antidote. But depending on the type of investigation, different tools may be used to reduce its effect. In a lab, keeping irrelevant information, such as whether the suspect has a prior record or confessed, from the analyst to avoid influencing judgment can solve any potential problem; and in a criminal investigation, an independent fact-checker who plays “devil’s advocate” and challenges the working hypothesis can protect the investigator.
In the case of the Temple physicist, there is no way yet to know whether cognitive bias – the fear that Chinese-born scientists were engaged in espionage – “made” the investigators see what wasn’t there: the “pocket heater” used in superconductor research that was supposed to be kept secret. But the question should be asked – internally at the FBI and in the Justice Department, and more generally in police or prosecutors’ offices.
The more we know about the risk of seeing what isn’t there, the better we can reduce the risk of another Xiaoxing Xi being wrongfully charged.
This article originally appeared as an op-ed at Philly.com. Jules Epstein is a member of the National Commission on Forensic Science. The views expressed are his own.
