Customs and Border Protection has wasted several years in developing technologies that would reduce the manual burdens associated with detecting threats and contraband in vehicles and cargo screened by non-intrusive inspection (NII) systems as they enter the U.S. at ports of entry, congressional appropriators say in recent reports.
Now, Congress intends to see that CBP gets going on a plan for introducing artificial intelligence and machine learning (AI/ML) into the NII systems.
Once the agency’s budget for fiscal year 2023 is signed into law, CBP will provide monthly updates to Congress on its AI/ML plans, Senate appropriators say in a report accompanying their version of the FY ’23 funding bill for the Department of Homeland Security. The report was issued in late July by committee Democrats without the support of any Republicans.
Senate Appropriations Democrats point out that in fiscal years 2019 and 2020, Congress provided CBP with more than $600 million in NII funding to substantially increase the percentage of personal vehicles and trucks to be scanned by large-scale X-Ray systems as they enter the U.S. at land ports of entry.
“A portion of those funds were to be dedicated to fundamentally changing how inspections were conducted,” the proposed Senate report says. “Rather than relying solely on a CBP officer to manually operate inspections, the use of artificial intelligence and machine learning capabilities would detect anomalies earlier and faster, pushing notifications to CBP officers for review.”
The report notes that developing the AI/ML technology for the NII systems was expected to take time and the work was supposed to proceed in parallel to CBP’s NII efforts.
“Unfortunately, that work did not take place and CBP has lost years in development and execution of this capability,” the Senate Appropriations Democrats say.
The Senate report is in agreement with one issued by House Appropriators in June in their version of the FY ’23 DHS budget.
In addition to being slow to spend the hundreds of millions of dollars CBP was provided to purchase the NII systems, the agency “also failed to plan for, or invest in the integration of artificial intelligence, machine learning, and autonomy into program,” the House Appropriators said. “Instead, CBP’s strategy continues to depend on CBP Officers to review thousands of images manually to hunt for anomalies while also attempting to match the images with manifests; there is nothing novel about this approach.”
House appropriators also dinged CBP for removing a requirement to provide under-vehicle X-Ray inspections of vehicles entering the U.S. Combined with the lack of investment in AI/ML development, the committee directed “DHS leadership” to address these “inexcusable” shortcomings of CBP.
House appropriators also rescinded $30 million in FY ’22 NII funding and redirected the funding “to focus only on creating the innovations first planned over four years ago.” And before any of this funding can be spent, the committee wants a plan from CBP on how it will adopt “innovative processes and technologies to scan and catalogue images, freeing up scarce CBP officers to perform other needed tasks.”
Overall, for FY ’23, House appropriators recommend that CBP’s NII acquisition program receive $50 million, which is $50 million more than the agency requested. Senate appropriators recommend $39 million for program.
The Senate report also “encourages” CBP to work with the DHS Science and Technology Directorate and the private sector to research and develop enhancements to existing NII equipment to create mobile systems to screen large-cargo and other items.