Here are questions and answers on the new lie detector being deployed in Afghanistan this month by the U.S. Army.
The answers are based on documents and interviews with Defense Department officials.
What is it?
The Preliminary Credibility Assessment Screening System is a hand-held computer or personal digital assistant that attempts to measure stress to decide if a subject is telling the truth. Initially the "P" stood for "Portable," but it became "Preliminary" to emphasize the idea that the device shouldn't be used to make final decisions. Either way, it's PCASS, pronounced "PEA-kass."
What does it look like?
The rugged TDS Ranger hand-held computer, which is about the size of a one-pound bag of coffee, is connected by a USB cable to a black wrist cuff. From there, two wires are connected to electrodes wrapped around the fingertips with Velcro or stuck to the palm; these measure changes in electrical conductivity of the skin. A third wire connects to a photo-plethysmograph, or pulse oximeter, which clips onto a fingertip and measures the interval between heartbeats. The wrist cuff converts the analog signals to digital and transmits them to the computer.
How does it work?
Like the polygraph, the PCASS uses external physiological information collected during an interview in an attempt to detect deception. These signals are interpreted by an algorithm, a computer program that displays the word "Green" if the person is thought to be telling the truth, "Red" for deceptive, and "Yellow" for an uncertain result.
Is the interview in English?
When the interviewee does not speak English, a translator is used for a pre-test interview. The questions are discussed, so they won't provoke a surprise reaction.
During the actual exam, the interpreter will read the questions from the screen. The operator punches in the answer, so the operator needs to know only two words of the language: yes and no. After a 25-second delay for gathering baseline information, the interpreter asks the next question. The entire process takes about 40 minutes.
Where is it being used?
It will be deployed this month in Afghanistan, then presumably in Iraq and elsewhere. The first customer is the U.S. Army, which has bought 94 units so far, but other branches of the U.S. armed services have expressed interest as well. Another 40 units have been purchased for training classes.
Who will it be used on?
It can't be used on U.S. citizens, according to the Pentagon's rules. (U.S. citizens in sensitive positions or under scrutiny are already eligible for a full polygraph, which is presumed to be more accurate than the PCASS; how much more accurate is a matter of dispute.)
When will it be used?
It could be used in many scenarios: interrogations after an IED explosion or screening of militia or Taliban members from groups of locals who want to work as security guards or interrogators or in service jobs.
The Defense Department says it will be used only for "triage," to pare down large groups to determine who will receive further scrutiny, not for final decisions.
What happens to people who get a red screen?
They could be detained, interrogated, polygraphed, turned down for employment, barred from military facilities.
What happens to people who get a green light?
They could be excused from further scrutiny, or allowed access to military bases and secure areas.
How is it different from a polygraph?
The PCASS collects less information and eliminates the examiner's role in reaching a decision.
Both devices collect information on electrodermal response (changes in sweating) with the fingertip electrodes. And both collect information on pulse and blood flow, though in different ways: The polygraph uses a blood pressure cuff, and the PCASS uses a pulse oximeter.
The polygraph also measures respiration with two sensors on the chest. The PCASS does not.
And the polygraph usually has one or more devices to detect countermeasures, or attempts to throw off the machine. A polygraph typically has a pad that the interviewee sits on, detecting movement such as tensing of the back muscles. The PCASS does not.
How are the examiners different?
A polygraph examiner for the Defense Department must be at least 25 years old, a graduate of a four-year college, with two years of investigative experience in law enforcement, and must complete a 13-week polygraph course and a six-month internship. The polygraph examiner analyzes information collected by the polygraph.
A PCASS examiner gets one week of training, either in the field or at the Defense Academy for Credibility Assessment. The PCASS examiner does not analyze the results; the device scores the exam.
Who sponsored its creation?
The work was supervised by the Defense Academy for Credibility Assessment, or DACA, formerly called the Polygraph Institute. Its office is at Fort Jackson, S.C.
DACA reports to CIFA, the Counterintelligence Field Activity, a secretive Pentagon unit whose mission is to protect U.S. forces against espionage, sabotage or assassinations.
The device itself is unclassified, and military officials agreed to show it to msnbc.com after the news organization had obtained studies describing the device.
CIFA itself may soon be shut down and its work folded into other defense agencies, the New York Times reported last week. The agency was created in 2002 when Donald H. Rumsfeld, then the defense secretary, moved more intelligence activities under the Pentagon after the Sept. 11 terrorist attacks. CIFA has had several public embarrassments, most notably the revelation that it maintained files on anti-war protesters in the U.S.
Who designed it?
The algorithm — which interprets the physiological information — was designed at Johns Hopkins University by the Advanced Physics Laboratory, where the project was called "Truth Verifier." That lab has done polygraph studies and was criticized for a lack of transparency in a 2003 National Academy of Sciences study of the polygraph's reliability. Johns Hopkins has been paid $1,194,000 thus far by the Defense Department.
Who sells it?
The finished device is sold by Lafayette Instrument Co. of Lafayette, Ind., which sells polygraph equipment. The cost is approximately $7,526 apiece, plus about $600 per year for maintenance and technical support. So far the Army has bought 94, and CIFA/DACA bought 40 for testing and training. The Defense Department says it doesn't have an estimate of how many will be purchased by the armed services for field use.
How accurate is it?
In Pentagon tests with basic trainees and civilians, the device was correct in 62, 63 and 79 percent of the cases. The Pentagon touts a higher number; setting aside cases where the machine couldn't decide, the accuracy was 74, 87 and 92 percent.
The device has had no scientific tests with data collected by the device near the battlefield.
Were there field trials?
The device underwent several "field vetting" tests in 2007, which the Defense Department says were intended to make sure the device was easy to operate and would stand up to field conditions. These were not scientific tests. There was no follow-up to determine whether the machine's positive and negative decisions were correct.
- Qatar: At CENTCOM's forward base in Camp As Sayliyah, the device was tested on employees of an Army maintenance contractor, from the Phillipines, Egypt, Nigeria, India and Nepal. They were asked about prior involvement with insurgents, militia groups and non-U.S. intelligence services. The first tests had an interpreter, but it was found out that he had a prior relationship with many of the examinees. Later tests were done in English, but many of the employees had poor English, so the test was abandoned. Of 39 tests, 21 were green, five red, and 13 yellow.
- Iraq: For access to a U.S. military compound, members of armed forces from Iraq, Egypt, Jordan, Sudan and the Philippines were asked about involvement with anti-coalition forces. Some tests were done in English; most in Arabic. Of 96 tests, 38 were green, 40 red, and 18 yellow.
- Camp Cropper: Iraqi correction officers and interpreters supporting an interrogation center near Baghdad International Airport were asked about involvement with anti-coalition forces and passing of information off post. Tests were in Arabic. Of 72 tests, 30 were green, 24 red, 13 yellow.
- Numaniya National Police Academy: Senior officers from the Iraqi police were asked about attacks against coalition forces. Of 28 tests in Arabic, 14 were green, seven red, seven yellow.
- Forward Operating Base Central, Iraq: Interviews in Arabic of counterintelligence special forces from the Iraqi Army. Of 11 tests, six were green, three red, two yellow.
What are the rules for its use?
The rules, set down in October by the under secretary of defense for intelligence, include:
- It cannot be used on U.S. persons.
- Each service must have a PCASS program manager and written procedures.
- Only certified personnel may use it.
- The first 25 interviews by an examiner must be supervised.
- Voluntary consent is required.
- Exam reports are kept for at least 20 years.
Will we see a commercial version?
Not from the Pentagon, but it wouldn't be difficult to construct a civilian version, at least as a novelty item. The components are easy to obtain: two electrodes, a pulse-oximeter, a box to convert their analog signals to digital. If you didn't insist on portability, you could skip the Ranger hand-held PDA, using instead a laptop or desktop computer. But you'd have to write the software, or modify commercially available software for the polygraph; the polygraph is expecting more information than the PCASS provides.
What's next?
Other lie detectors are on the way from the Defense Department.
On a tour of the lab at the Defense Academy for Credibility Assessment, research director Troy Brown showed off a series of devices being tested:
- A thermal facial imager (known informally as "the booth of truth"), to measure changes in heat in different areas of the face.
- A laser that can be pointed at the carotid artery to detect stress.
- Eye scanners that can tell where a person is looking on a computer screen, perhaps giving away guilty knowledge.
The credibility assessment team is also supporting university research on functional magnetic resonance imaging, or fMRI, which looks at brain activity. Although companies are already selling brain scan services for lie detection, under names like No Lie MRI, Brown pronounces each of these research projects to be years, probably decades, from valid use.