Treasury inspector general warns of errors in IRS sharing taxpayer info with ICE
Happy Tuesday. Here’s your Tuesday Tech Drop, the past week’s top stories from the intersection of technology and politics.
Treasury insecurity
The Trump administration’s choice to deepen the Treasury Department’s involvement in its racist anti-immigrant crackdown may have exposed taxpayers’ personal information, according to a new report from the Treasury’s inspector general.
The IG’s report “raises concerns about Immigration and Customs Enforcement’s ability to safeguard taxpayer information,” The Associated Press reported Monday. The IG report found the administration’s arrangement for the IRS to share data with ICE was error-prone and unreliable.
Under this arrangement, ICE can submit the names and addresses of people the agency says are in the United States illegally, and the IRS cross-references that information with tax records.
The arrangement has been challenged by lawsuits and led former acting IRS Commissioner Melanie Krause to resign from her role in February.
Any errors that result from this information sharing are important to spotlight, as the Trump administration increasingly looks to the Treasury for help with its immigration crackdown.
According to the AP:
The report states that after the agreement was signed, ICE requested address information on more than 1.2 million people, and the IRS ultimately provided last-known addresses for about 47,000 people. [The inspector general] concluded that the IRS’s automated matching process was flawed. Inconsistent formatting in ICE’s data led to questionable matches, including in cases where incomplete or inaccurate addresses were labeled as valid, the report says. Representatives from Treasury and the IRS did not respond to an Associated Press request for comment.
Read more in the AP.
New York turns to AI for help with food aid
New York City is using an artificial intelligence tool developed with help from consulting firm McKinsey & Company to administer SNAP benefits. The tool will reportedly be used to identify payment errors, as the Trump administration has vowed to withhold vital food aid funding from states with higher error rates. Given AI ethicists’ concerns about algorithms meting out discrimination, this tool certainly requires oversight to ensure it doesn’t facilitate harm.
Read more in the New York Daily News.
Miller’s spy fantasies spook conservatives
Some prominent conservative influencers and lawmakers had a conniption after White House deputy chief of staff Stephen Miller publicly objected to efforts to limit the Trump administration’s spy powers.
Read my report on the backlash to Miller on MS NOW.
Trump’s AI orders for the military
President Donald Trump, who’s reportedly made investments in AI companies and whose sons are linked to drone companies, issued an executive order Friday calling on the military to accelerate its use of AI.
Read more in the AP.
Trump eyes U.S. stake in AI companies
Speaking of Trump and AI, my colleague Steve Benen wrote about the president’s proposal for the government to take a stake in AI, continuing his quasisocialist (arguably, authoritarian) trend of having the government take a stake in supposedly free-market enterprises.
Read more on MS NOW.
Meta’s security problem
Meta spokesperson Andy Stone said the social media company fixed a security issue that allowed hackers to use its AI chatbot to access Instagram users’ account passwords.
Read more at Tech Crunch.
More Meta, more (alleged) problems
A new report from the Center for Countering Digital Hate, a digital watchdog group, found that violent threats and hate speech toward lawmakers spiked last year after Meta rolled back some content moderation policies designed to protect users from harassment, bullying and threats of violence. After Meta made its decision, human rights groups warned the move risked fueling violence. Last year, Meta’s oversight board issued a ruling that found the company acted “hastily” and with little regard for global impact when it made those content moderation changes.
In a statement to MS NOW, a spokesperson for Meta said:
We regularly issue public reports tracking violating content on our platforms, and the prevalence of hateful conduct did not increase throughout 2025. We cannot address the claims in this report as we were not provided it in advance of publication.
Read more on the CCDH’s “Safety Off” report here.
Santos scandal saga continues
My colleague Erum Salam got the scoop on a federal insider trading probe related to alleged online betting by scandal-plagued former Republican New York Rep. George Santos, whose prison sentence for fraud was commuted by Trump last year. Santos called the insider trading claims “preposterous” in a social media post.
Read more on MS NOW.
The post Treasury inspector general warns of errors in IRS sharing taxpayer info with ICE appeared first on MS NOW.
