This year, the FBI will roll out a national Rapid DNA network with police using portable DNA testing in a “magic box” the size of a desktop printer. These magic boxes can analyze a swab, produce a profile in about two hours, and query the local or federal DNA data banks.

The federal government is aggressively supporting this effort through funding and legislation as it hopes to expand the coverage of its national DNA databank. The appeal of speedy DNA results is completely understandable, particularly where crime labs can be backlogged by many months.

Yet as recent coverage of the police rollout of these devices makes clear, the advent of Rapid DNA introduces a host of concerns. We lack much meaningful oversight of forensic science or police technology in this country. These machines are unvalidated. They use proprietary “black box” technology, so that no one, least of all the people using them, knows precisely how they work. They make errors and destroy samples, but we do not know how often.

And yet they are still being introduced at police stations around the country. Quick and dirty forensic tests will inevitably cause criminal cases to go terribly wrong. Innocent people may be convicted.

Additionally, the police officers operating these magic boxes do not necessarily get any training on how to collect DNA or test it properly. They are not trained scientists working at a crime lab, and may not be operating under any quality controls. “I barely need a pulse to use this instrument,” a detective in Bensalem, Pennsylvania, told The New York Times. He explained how he had just a few hours of training from the manufacturer.

While testing a single swab from a person can be simple, these machines cannot interpret crime scene DNA samples, which can contain DNA mixtures from an unknown number of people. The few studies done of these machines have suggested high error rates, including the capability to destroy samples.

We have seen this time and time again when new portable technology is rolled out without validation or regulation. Take the story of a Tampa Bay, Florida, woman who had run out of gas. When a police officer arrived, he saw pills in the car and suspected drugs, even though she insisted they were vitamins. He conducted a field test using a chemical kit, which identified them as oxycodone pills. She was arrested and charged with drug trafficking, languishing in jail for five months before the state crime lab did additional tests and found they were in fact vitamins.

Hers was far from an isolated case. Hundreds of wrongful convictions have come to light around the country involving field drug tests. In Harris County, Texas, an audit by the prosecutor’s Conviction Integrity Unit uncovered 456 cases in which field drug tests gave erroneous results, and the convictions in those cases were all reversed. In 2016, the Texas Forensic Science Commission said these field tests are too unreliable to use in criminal cases and should always be followed up with lab tests.

Many police departments are doing just that with Rapid DNA — conducting a field test, but then sending evidence to a crime lab for a more involved check. Other police may not do so. And during any delay at the crime lab, innocent people may plead guilty.

As we extend these capabilities to police departments and outside the lab, we cannot escape how poorly regulated forensics are in this country. Indeed, the National District Attorney’s Association has said it opposes the use of Rapid DNA machines except by experienced DNA analysts at accredited crime labs. Absent rules to validate magic boxes or a commission like the one in Texas, who will step in and put a stop to error-prone technology being used by untrained people?

We have seen this all happen before. Unless new forensic technology is carefully validated and regulated, the result will be lost evidence, the guilty going free, and innocent people getting jailed. Some will plead guilty, some may eventually get cleared in the courts, and many will never have their stories told.

Brandon Garrett is the L. Neil Williams Professor of Law at Duke University School of Law. He is the author, most recently, of “End of its Rope: How Killing the Death Penalty Can Revive Criminal Justice.”