Last weekās claims were false things built on real facts. This weekās are harder to handle, because none of them is false. Each one is a real event with a single word attached that the event did not earn: a ruling became a purge, a poll became an all-time low, a repeated sentence became a humiliation, and an intelligence assessment about āChinese groupsā became China. The job this week is to find the word, take it off, and see what is left standing, because in every case what is left is worth more than the version that traveled.
The voter purge the Supreme Court just approved
The claim, from both directions at once. On the left, since Friday: the Supreme Court has cleared Trump to purge noncitizens from the voter rolls before the midterms. NBCās headline said the Court allowed the administration to use the database āfor potential voter purges,ā and smaller outlets dropped the word potential. On the right, the same ruling ran as the Court āgranting states the ability to verify voter status in the midterm election,ā and the Department of Homeland Security put out a release applauding it.
What happened when it was checked. On Friday the Court issued an unsigned order pausing a June lower-court ruling that had blocked the expanded SAVE system, with Justice Jackson dissenting. That is a stay rather than a decision on the merits; the case continues on appeal. SAVE is the Systematic Alien Verification for Entitlements database, built to check eligibility for public benefits, which the administration expanded in 2025 so election officials could run entire voter rolls against it in bulk, including by Social Security number. What the ruling does is let DHS switch that back on while the litigation runs.
What it cannot do is produce a purge before November, because federal law already forbids one. The National Voter Registration Act bars systematic removal of voters from the rolls within 90 days of a federal election. The midterms are November 3. The 90-day window closed in early August. Officials can check an individual voterās citizenship in that window, and they can queue up bulk results for later, but the mass removal both sides described cannot legally happen until after the polls close.
What was true underneath. The toolās error rate, which is the part neither headline mentioned. ProPublica found the expanded SAVE system flagging citizens as noncitizens in at least five states. In Texas it flagged hundreds of people who had registered through the Department of Public Safety, which requires proof of citizenship to register in the first place. The rulingās real effect lands in 2027 and 2028, when the quiet period lifts and states can act on bulk matches. That is when the false-positive rate becomes the whole story, and it is a better use of attention now than a purge that the calendar rules out.



