Instead of paying first and investigating later, he said Congress should build systems capable of stopping fraudulent payments before they leave Treasury. "We have access to technology today so the fraudulent checks never, ever, ever go out the door," he said. "This shouldn't be partisan, but it is technology. We need to take that on." ... "We're going to have to also take on pieces of legislation that we've passed and stop the scammers from scamming us."
Build prevention-first federal payment systems that use modern technology to stop fraudulent payments before taxpayer money leaves the Treasury and revisit program designs that enable waste and fraud.
Occurrences
Evidence
H.R. 7972 was introduced by Mr. Schweikert and would require an IRS fellowship program for qualified data scientists, advanced analytics, statistical modeling, machine learning, AI/data-analytics use-case review, anomaly detection, and recommendations for improving improper tax payments.
Schweikert's office said the Ways and Means Committee passed H.R. 7972 to modernize IRS operations with private-sector data science expertise, advanced analytics, statistical modeling, machine learning, and improper payment prevention.
The Ways and Means markup record says H.R. 7972 was ordered favorably reported to the House, as amended, by a 24-16 roll-call vote on July 1, 2026.
H.R. 9501 lists Mr. Schweikert as an original cosponsor and would require the Treasury Secretary to establish an AI pilot program to identify inaccurate tax returns caused by identity theft, fraudulent refund claims, or improper preparers.
JCT described present law as having no IRS program to use AI to identify inaccurate tax returns, and said H.R. 9501's proposal would take effect only on enactment.
Treasury's Do Not Pay program says it gives agencies data and services to verify recipient identity and eligibility before awards or payments, and helped prevent, detect, and recover $11.7 billion in potential fraud and improper payments in FY2025.
Assessments
Schweikert materially advanced the promise through H.R. 7972, which he introduced and moved through Ways and Means, and by cosponsoring H.R. 9501 on AI tax-fraud detection. But the evidence shows these were proposed or committee-reported measures, not enacted law or implemented federal payment systems. Treasury's Do Not Pay program already existed and does not establish that Schweikert delivered the promised new prevention-first system or program redesign. This is a serious same-term legislative effort, but the promised outcome was not fulfilled.