My colleague David Zimmer dug into the data to assess the success of President Trump’s immigration policies. He found them to be very successful, indeed:
There has been a concerted effort by “open immigration” advocates, “anti-Trumpers,” and a supportive legacy media, to portray the immigration enforcement efforts of the Trump administration in the poorest light possible.
Many detractors have virtually ignored the success the Trump administration has had at securing the border, while others have suggested the administration has not held to its “stated goal” of focusing on the “worst of the worst.” A common tactic has been to use selective data and narrow parameters to suggest that only a “small” proportion of immigration arrests have involved people with criminal histories – a narrative that simply doesn’t withstand scrutiny.
The data follow, including these numbers:
FYTD – 2026: 353,948 individuals have been booked into ICE detention facilities.
* 60% have had a criminal history or a pending criminal charge
* 82% have been “released for removal”
* 0.4% have been “paroled” into the US
* 0.11% of those booked with a criminal history or a pending criminal charge have been “paroled” into the USCustoms and Border Protection facilities have experienced a 167% decrease in the daily average number of detainees compared to FY 2024.
Conclusion:
During its first 18 months the Trump administration has undeniably secured the US border – while dramatically increasing the number of illegal aliens detained within the US (including a larger percentage of detainees who are criminal offenders), and decreasing the number of illegal aliens paroled into the US as compared to the final year of the Biden administration.
A very positive picture. But here is the kicker. David asked Google’s Gemini AI program what it thought of the data he had assembled:
I thought I’d put AI to the test to see what its response would be regarding the data I presented above. It responded twice that I was misrepresenting the data relevant to both the number and percentage of illegal aliens detained who had either a criminal history or pending criminal charge. AI referenced the CATO Institute and the Migration Policy Institute as source documents for refuting my conclusions. I then provided the link to the DHS raw data and asked AI to calculate the figures and get back to me. Here’s AI’s response – I’ll take the win! Just an example of the bias in our world, and how that shapes what we are told is the “truth.”
“You are entirely correct to call this out, and I owe you a direct apology. In my initial answers, I failed to look at the raw data directly from the ICE Detention Management Dashboard, relying instead on standard summaries and external policy institute narratives. This was a mistake on my part, and your push for the raw math was completely justified.”
The apology is welcome, but it would be better if AI programs didn’t regurgitate liberal talking points in the first place.
