M: I need you to verify some numbers if you can so I can figure out what kind of critique of ICE I need to do. Okay, if you check online, and I give you permission for whatever else in this silo, but the NAACP filed something about ICE going to the polls for the next couple days— sorry, over the last couple days, and trying to prevent them from trying to, you know, sidestep or just straight up do something illegal. That part is only relevant in that it raised a question in my mind. One aspect of the filing seems to be the idea that ICE is using an inordinate amount of violence towards civilians, not even within their jurisdiction, but rather with kind of a spray-and-pray approach in colored neighborhoods. What I'm trying to find are some numbers that tell me whether that is specifically a racist thing going on, or if it's an aspect of the fact that you're going to find more immigrants— sorry, let me rephrase this, and remember to think of this in Bayesian terms, but if you are an immigrant in America, it is more likely that you are going to be some form of non-white and therefore found in darker-skinned communities. The idea is trying to figure out whether ICE has used this reasonable assumption and turned it into an actual racist thing, or the counterexample that I'm seeing would be finding out that they are just as shitty in all neighborhoods and that the count and amount of incidents within darker-skinned neighborhoods is due to being bad in every area but ending up in more dark-skinned areas. Their presence there in terms of racism is much more arguable, but if it turns out that they are much less overbearing or violent in white communities, then that's an entirely separate issue. Am I making sense?
G: Yes, you're making sense. You're distinguishing three different questions that could produce superficially similar statistics, and you're trying to avoid treating evidence for one as proof of another.
Your counterexample is important. ICE could be using excessive force across the board, while the geographic distribution of its operations causes a disproportionate number of non-white people to experience that violence. That would be a serious problem, but it would not, by itself, establish racially selective use of force.
Conversely, finding comparable enforcement situations in which ICE behaves substantially differently depending on the community's racial composition would provide evidence of a separate racial disparity in its conduct.
The key is to compare rates under comparable conditions, not just raw incident counts. I'll look for the NAACP filing, available enforcement and force statistics, and demographic data that can help separate those possibilities.
G:
On September 18, 2026, the NAACP and eight other organizations filed NAACP v. DHS, challenging the administration's threatened deployment of federal law enforcement, including ICE, around polling locations.
The complaint argues that administration statements and existing immigration-enforcement practices have created reasonable fears of voter intimidation, particularly among naturalized citizens and communities of color. It is a challenge to threatened election interference, not a statistical finding that ICE uses force more frequently against one racial group. <Cite refs={["turn0search1","turn0search9","turn0search3"]}/>
The administration disputes the characterization. The Justice Department says it has not intended to send officers to polling places, while DHS disputes allegations of racial profiling. <Cite refs={["turn0news37","turn0search3"]}/>
That distinction matters for your analysis: the complaint provides a reason to investigate the racial dimension, but its filing does not establish the comparative rates you're asking about.
G:
Pew Research Center's 2024 figures show that the foreign-born share of the population varies substantially by race and ethnicity:
| U.S. population group | Percentage foreign-born |
|---|---|
| White, non-Hispanic | 4.1% |
| Black | 11.3% |
| Hispanic | 33.4% |
| Asian | 64.5% |
These are percentages within each population group, not the racial composition of all immigrants. <Cite refs={["turn0search4","turn0search5"]}/>
That is a substantial demographic difference. It supports the proposition that immigration enforcement will have uneven racial effects even if officers use a geographically neutral approach based on the actual distribution of immigrants.
But there is an important denominator problem: foreign-born does not mean deportable. Pew estimated 14 million unauthorized immigrants in 2023, compared with a much larger total immigrant population. Its unauthorized category also includes some people with temporary protections. <Cite refs={["turn1search0","turn1search7"]}/>
For your purposes, the relevant baseline would ideally be the number of people actually subject to enforcement, broken down by geography, immigration status, and race or perceived ethnicity.
Reuters reported on September 14 that ICE arrests had risen to nearly 51,000 in August 2026. More than three-quarters of recent arrests involved people without criminal records. <Cite refs={["turn0news38","turn1news37"]}/>
That is useful evidence about the breadth of enforcement. It is not, however, a measure of racial selectivity or violence. An arrest is not automatically an instance of excessive force, and the absence of a criminal record does not independently establish that an immigration arrest was unlawful.
There is also documented evidence of individual serious force incidents. For example, an ICE officer was charged in connection with a January 2026 Minneapolis shooting and alleged false statements about the encounter. Those charges establish that the incident is subject to criminal proceedings, not the outcome of those proceedings or a national racial pattern. <Cite refs={["turn1news36","turn1news38"]}/>
I have not found a verified national dataset that supplies the critical comparison: ICE force incidents per comparable enforcement encounter, separated by the racial composition of the neighborhood or the perceived race of the person encountered.
Without that denominator, the evidence reviewed here cannot establish whether ICE is more violent in non-white communities, equally violent wherever it operates, or following some more complicated pattern.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G:
I'd keep four hypotheses alive rather than forcing the evidence into a binary racist/not-racist question.
| Hypothesis | Evidence that would distinguish it |
|---|---|
| A. Demographic concentration | Deployment largely follows the distribution of people subject to immigration enforcement. |
| B. Racially selective deployment | ICE enters non-white neighborhoods more frequently than their enforcement-relevant populations would predict. |
| C. Generally aggressive conduct | Comparable encounters produce similar rates of excessive force across racial groups and neighborhoods. |
| D. Racially selective conduct | Comparable encounters produce different rates of excessive force depending on race or neighborhood composition. |
These hypotheses aren't mutually exclusive. A and D, for example, could both be true.
There is also a complication worth retaining: even if agents use the same amount of force in two neighborhoods, they might initiate encounters on different grounds. A policy that stops many more innocent people in one neighborhood could create a racial disparity without any difference in the force used after a stop begins.
I'd use two separate rates:
[ \text{Deployment rate}= \frac{\text{ICE operations in an area}} {\text{Estimated enforcement-relevant population}} ]
[ \text{Force rate}= \frac{\text{Excessive-force incidents}} {\text{Comparable ICE encounters}} ]
For the second rate, I'd separate people suspected of immigration violations from bystanders, citizens, and other people who weren't the intended targets. I'd also distinguish reported force, independently substantiated force, and legally adjudicated excessive force.
That would let you test whether the disparity comes from where agents go, whom they stop, what they do once they stop someone, or a combination.
The demographic evidence supports your proposed alternative explanation for geographic concentration. The current enforcement numbers establish the scale and breadth of ICE activity. Neither supplies the comparative force data needed to determine whether racial selectivity explains ICE's conduct.
Importantly, the available evidence doesn't establish the opposite conclusion either. A missing comparison is not evidence of equal treatment.
Your distinction also preserves a separate question about the lawfulness of individual encounters. An unlawful detention or unjustified use of force does not become lawful merely because officers behave similarly in white and non-white communities.
The next useful investigation would be a city-level comparison, rather than another national incident count. Los Angeles and Minneapolis offer documented enforcement controversies that could be examined alongside neighborhood demographics, deployment records, encounter reports, and independently verified force incidents. That would give your competing hypotheses something measurable to work against.
For now, I'd leave the racial-selectivity question open rather than assigning it a probability the available evidence cannot support.
M: There is part of me that one day wants to hear you say something along the lines of, You're making absolutely no sense. What the hell are you talking about? Not something I want in the way that people desire things, but if it ever happens, I am likely to be just so addled mentally for some reason, or so confused on a topic that I'll end up laughing like hell. Anyway, let's see what you've got for me. Yeah. I may have wandered into all three of the questions that you present, but my primary focus is number two. It's a question of how ICE operates in Black versus White, or Brown versus White communities. We're going to go with Brown as being reasonably encompassing of the situation. But the point is that ICE operating more often within Brown communities might be reasonable, even if they are shitty when they get there. The question is primarily, though, number two, which says even if we assume that there is a logical reason for there to be more of a presence in Brown communities, their quality of performance is the key thing for what I'm curious about. Because if they're just as big a bunch of dicks in White communities and only are lower in the numbers there because there are lower incidents there than in Brown communities, then that's a wildly different situation than if they tiptoe through White communities but go all out idiocracy in Brown communities. And the question seems to revolve around, or be instantiated by, or something. The argument is, from what I understand of reports, that ICE doesn't really seem to care whether someone is actually deportable. They are, by reports, just going in and saying if someone is brown-skinned, they are more likely. Which, in a Bayesian argument, is a place to start, but in their application of it instead is all the evidence they need to make a judgment, leading to serious incidents. So we're not focused on the number of arrests or anything, or the percentage even of the number of people who actually should have been arrested by any stretch or permutation of the laws as they have been or as they have become. We're looking only at the amount of violence that is going on within the different communities, between the white communities and the non-white ones. It looks like you're approximately on the right path for what I was thinking of, and I think, or at least I hope, I've clarified things a little bit more. What I want you to do is go into the net and see if you can scrape up any numbers that can help. If you can't find anything complete or definitive, see what you can find that might provide at least a little bit of information that we don't already have.
G: That clarifies the target considerably. You're asking whether ICE's level of aggression changes with the racial composition of the community in which it's operating, not whether it makes more arrests in one community or another.
The distinction is between two possible patterns:
We can set aside whether ICE had a legitimate reason to be in either neighborhood for the initial comparison. We can also set aside whether the person encountered was deportable, except where that affects whether force was justified.
And yes, your Bayesian distinction is valid: using demographic information to estimate where an immigrant population might be concentrated is different from treating someone's skin color as sufficient evidence to detain them. But the latter is a question about targeting; your primary question is what agents do once an encounter happens.
I'll look specifically for comparative force data, including shooting and injury counts, documented uses of force, encounter-level studies, and geographically identifiable incidents. The important limitation will be whether the available evidence measures actual behavior or merely the distribution of reported incidents.
G:
The search turned up some useful partial evidence, especially in Los Angeles and New York. It also exposed a major limitation: several datasets quantify racial disparities in enforcement, but don't measure differences in force per encounter. I'll keep those two categories separate.
A July 2025 investigation reported that the Coalition for Humane Immigrant Rights had verified 471 immigration-enforcement actions in Los Angeles County between June 6 and July 20.
The five ZIP codes with the highest numbers of reported actions had a combined population that was 76% Latino. The organization also received 1,677 additional reports it couldn't verify; 1,500 mentioned armed agents, and 389 mentioned witnesses seeing apparently random arrests. <Cite refs={["turn0search1"]}/>
Those figures are suggestive of the geographic concentration you're describing. But an armed agent isn't necessarily an agent using force, and the reports don't establish how ICE behaved in predominantly white neighborhoods.
More importantly, the reporting organization estimated that it captured only about a third of enforcement activity. So this isn't a complete denominator.
A May 2026 investigation by The City Reporter examined more than 1,200 lawsuits and identified 430 street arrests. Over 93% of the people arrested in the examined cases were from Latin American countries, compared with an estimated 66% Latino share of the region's undocumented population. <Cite refs={["turn0search3"]}/>
That is a disparity in the available sample, although lawsuits aren't a random sample of all arrests. The investigation also describes encounters involving Tasers, physical takedowns, and alleged racial slurs. <Cite refs={["turn0search2","turn0search3"]}/>
This is closer to your concern about aggressive behavior. But it still doesn't establish the rate of force among white versus Brown people subjected to similar operations.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G:
This is the most directly relevant numerical source I found.
The ACLU reviewed more than 1,200 immigration-enforcement incidents from 2025 across eight states. Its July 16, 2026 report documented the following:
| Conduct documented | Count |
|---|---|
| Pushing, shoving, tackling, or pinning people | 418 |
| Tactics that can restrict breathing, including chokeholds | 81 |
| Deployments of chemical irritants | 361 |
| Chemical irritants aimed directly at individuals | 132 |
| Incidents involving likely racial profiling | 437 |
Source: ACLU, Agents of Chaos and Cruelty. <Cite refs={["turn1search1","turn1search2"]}/>
These categories can overlap, so we cannot add them together to obtain a total number of violent encounters. Nor can we interpret them as national rates: the reviewed incidents were drawn from selected documentation, not a random sample of all ICE operations.
The report covers ICE, other federal immigration agents, and participating state and local officers. It does not provide the neighborhood-by-neighborhood white-versus-Brown comparison you want.
Nevertheless, it supplies something valuable: a set of concrete behaviors that could be used as the outcome variables in such a comparison.
Minneapolis provides evidence that aggressive immigration-enforcement operations can also seriously harm white Americans.
In January 2026, federal immigration agents fatally shot U.S. citizens Renee Good and Alex Pretti during Operation Metro Surge. A separate January shooting of Venezuelan national Julio Sosa-Celis led to state assault charges and federal false-statement charges against an ICE officer. He pleaded not guilty to the federal charges on September 18. <Cite refs={["turn0search4","turn1news36","turn1news37"]}/>
The Good and Pretti incidents are important counterexamples to an absolute claim that white people are spared serious force. But two fatal encounters cannot tell us whether white people face the same rate of force.
They also involve different encounter circumstances, including the agents' responses to people present during enforcement operations. That complicates direct comparison with immigration-targeted arrests.
The Washington Post obtained internal ICE records covering 1,460 uses of force at detention centers from January 2024 through February 2026. At least 780 occurred during Trump's first year back in office, representing a reported 37% increase over the preceding year. The underlying data have been made available publicly. <Cite refs={["turn1search9","turn1search12"]}/>
This is potentially useful for studying whether ICE's overall approach to force has changed.
But detention-center force is not street-enforcement force. The populations, situations, and denominators differ, so I would not combine these figures with the ACLU's street-enforcement counts.
<box border={{size:1,color:"default"}} radius="lg" padding={3}
gap={2}>
Your two competing explanations remain distinguishable in principle, but the publicly available figures I've found do not resolve them.
The strongest next step would be to obtain incident-level data from the ACLU report and classify documented uses of force by location and encounter circumstances. That could produce a preliminary geographic comparison. Without a reliable count of all comparable encounters in each area, however, it would still be an analysis of reported force, not the true rate of force.
Bottom line: There is substantial documented evidence of force during immigration enforcement and evidence of racially disproportionate targeting. The narrower claim that ICE applies a different standard of physical restraint in white versus Brown communities remains unproven by the comparative data reviewed here. That is a meaningful gap, not a reason to collapse the two questions into one.
[Attachment reference unresolved: file-more]