Background-Check Technology Advances, but Concerns Linger

​As the demand for online ordering and delivery services grows, companies that use drivers to transport people, packages, parcels, food and merchandise are depending more on artificial intelligence to perform criminal background checks before drivers are hired.

One company that relies on background-check technology is Delivery Drivers, Inc. (DDI), which is based in Irvine, Calif., and is a third-party administrator for independent contractors. The company recruits, onboards, screens and enrolls drivers for clients in the courier, restaurant delivery, retail and grocery delivery businesses.

"As a third-party administrator in the 1099 independent contractor space, we conduct background checks to ensure compliance with 1099 regulations," said Carly Fliesher, director of strategic partners at DDI.

The COVID-19 pandemic has significantly increased the demand for delivery services. Overall employment of delivery drivers is projected to grow 12 percent from 2020 to 2030, faster than the average for all occupations, data from the U.S. Bureau of Labor Statistics shows.

The shutdown of retail stores at the beginning of the pandemic increased online shopping on sites like Amazon, which in turn grew the demand for drivers delivering goods to homes and offices. Companies like DoorDash, Uber Eats and Grubhub saw their revenues surge as online food ordering became more popular.

With all of these new employees comes the need for background checks. DDI has worked with Checkr Inc., a San Francisco company that handles pre-employment screening and background checks on job applicants, for the past five years.

DDI currently has 50,000 active contractors. The company uses two different types of Checkr services: motor vehicle records to verify the age of the job applicant, and background checks to look for past misdemeanors or felonies that the applicant may have committed.

"In the past, customers cared more about motor carrier records, but now safety is becoming more of a concern for our clients," Fliesher said. "They want background checks because customers don't want someone who has stolen something or has committed identity theft to be handling credit cards."

She added that Checkr's machine learning tool has helped DDI realize a 300 percent improvement on the time spent on the adjudication process. The company averages about 7,900 background checks a week for clients that are increasingly concerned about new hires with criminal records.

Advancements in cloud computing allow Checkr's platform to process larger volumes of data. Application programming interfaces (APIs) provide better integration between Checkr's tools and customers' systems, and the use of machine learning algorithms speeds up the time it takes to perform background checks.

"What used to take three to five days now takes one and a half days for a background check to be completed," Fliesher said.

There are several competitors that offer other screening services, including Accurate Background and First Advantage, which provide screening for international job candidates.

Kristen Faris, senior vice president of sales solutions at Checkr, said one of the common misconceptions is that the company uses AI to make decisions around hiring. Instead, the company leverages machine learning to normalize and standardize information gathered from court records across the United States.

"For instance, different courts have different ways of naming their criminal records. In one court, burglary might be called burglary, in another it might be called Forced Ent, and in another it might be called Burg," Faris said.

Through machine learning, Checkr standardizes criminal conviction nomenclature across all the 6,000 jurisdictions in the U.S.

"If DDI as an employer said we will not hire anybody with a felony burglary conviction, our system can automatically detect the felony burglary charges across the United States and then in an automated fashion apply the adjudication criteria to it," she said. "It's more about streamlining the process than it is about the actual [hiring] decision, which is still with the customer at the end of the day."

Making decisions based on accurate background checks is critical to Checkr, which has been bruised by legal challenges. In 2019, Checkr settled a class-action lawsuit that alleged the company illegally included information about low-level offenses like traffic infractions on background checks for more than 96,000 people. Checkr agreed to pay $4.46 million in damages to settle the lawsuit.

Mistakes in background-check information can cause considerable harm to a job applicant, said Merritt Maxim, vice president and research director at Cambridge, Mass.-based Forrester Research Inc. "With automation, if the tooling is not tuned correctly or if it is working with false data, it may generate inaccurate results, which could prevent the job applicant from being hired," Maxim said. "Correcting the court's data may involve talking to multiple government agencies, which takes time, and that means the potential impact of inaccurate background checks could be longer-term."

Responding to Checkr's legal challenges, Faris said the company's mission of building a fairer future is central to its background-check technology and that all candidates, regardless of their criminal history, should have a fair chance to work.

"At our core, Checkr is a company using technology that makes fair-chance hiring at scale possible," she said. "Checkr routinely audits its processes and technology (including technology leveraging machine learning) to ensure the accuracy of its systems, and we continue to improve our processes and technology to increase accuracy," Faris said.

Nicole Lewis is a freelance journalist based in Miami.