My Favorite Use Case for AI
The Best Sanity Check Tools
A recent social media post from a recommended account gave me whiplash. As a non-native English speaker, I’ve learned to run sanity checks when something makes me wonder whether I’m missing context or whether people really can make so little sense. This was one of those moments.
Indiana state Senator Mike Bohacek posted a statement on Facebook responding to Trump’s use of an ableist slur against Minnesota Governor Tim Walz. I initially saw only part of the post and consulted my thinking A.I.des to parse what seemed like incoherent logic. But when I found the full statement, it became clear this wasn’t just poorly reasoned—it was actively disturbing.
Here’s what he actually wrote:
I read it twice. Then a third time. Redistricting—a fundamental democratic process affecting how millions of Indianans will be represented for the next decade—has exactly zero logical connection to Trump’s social media language. Yet Bohacek explicitly links them: he’s voting NO on redistricting as retaliation for Trump’s offensive language, using his daughter as justification.
So I did what I always do when faced with something that doesn’t square—I consulted my thinking A.I.des to see if I was missing something or whether this really was as incoherent as it appeared.
What the AI Caught
The logical failures were immediately apparent to GPT-5.1 and Claude Sonnet 4.5, although Gemini Flash charitably viewed Bohacek’s remarks as effective political communication likely to appeal to his base, who might find politicians taking stances on issues where they have no personal stakes to be phonies. Gemini 3 Pro initially agreed but reversed its assessment after further analysis, providing a devastating summary that I’ll include below along with highlights from my other thinking A.I.des’ analyses.
The post offers no actual position on redistricting based on its merits despite opening with that question. It substitutes personal biography for principled argument. It explicitly treats a legislative vote as a personal weapon to punish someone for hurt feelings. Any undergraduate writing course—or apparently any LLM trained on basic argumentation—would flag this as structurally and ethically unsound.
But the deeper problem emerged through discussion: this isn’t just poor communication. It’s evidence of severely constrained moral reasoning combined with a willingness to sabotage democratic processes for personal grievances. That single sentence—“I have been an unapologetic advocate for people with intellectual disabilities since the birth of my second daughter”—reveals someone whose empathy only activates through direct personal experience.
Not since reading about disability rights. Not since meeting constituents with disabled children. Not since studying policy impacts. Since his daughter was born. And not for all disabilities—specifically intellectual disabilities, because that’s what affected his family. The scope is breathtakingly narrow.
Secondhand Experience Is a Perfectly Good Foundation for Sound Judgment and Knowledge
Here’s what strikes me as extraordinary: I’m working with language models that have zero firsthand experience of anything. They’ve never met a person with disabilities, never watched a child struggle, never felt personal grief or joy. Everything they know comes from linguistic patterns in training data—pure secondhand knowledge.
Much of my own knowledge that informs my judgment also has the same source. As entities with language faculties, we can all read and consume media that teach us about the world without having to experience every single thing in person. An older cousin once tried to convince me that firsthand experience was the only reliable foundation for grounded knowledge; even as a middle schooler, though, my love of crime fiction and biographies immediately reminded me of quite a few firsthand experiences I did not want to have and could sufficiently learn about through books and movies.
And yet we all can engage coherently with concepts like universal human dignity, systemic injustice, and principled policy-making in ways that Senator Bohacek apparently cannot. We can recognize that opposing ableist language shouldn’t require having a disabled family member, just as opposing racism shouldn’t require being a racial minority or opposing religious persecution shouldn’t require sharing someone’s faith. Every functioning democracy rests on the expectation that leaders can care about harms they have not personally suffered.
This example is an indictment of how poorly some humans use their capacity for abstract thought and empathy. Bohacek has decades of lived experience, access to constituent stories, policy research, historical precedent, and presumably some exposure to philosophical and religious frameworks about universal human worth. He has infinitely more resources for moral reasoning than this ESL speaker or any language model.
Yet his moral imagination apparently only extends to people who remind him of his daughter. That’s operating with less empathy than a statistical pattern-matching system, which seems theoretically impossible but is empirically demonstrated by that Facebook post.
The Selective Outrage Problem
The narrowness becomes even more disturbing when you zoom out. Bohacek found his voice over a slur directed at Tim Walz, who is not disabled. But where was this unapologetic advocacy in 2017 when children were sleeping on concrete floors in detention facilities, when families were systematically separated, and when tracking systems failed and some children remain lost years later?
Those were actual children experiencing actual harm—not insensitive language, but measurable trauma with documented long-term psychological effects. Yet that didn’t trigger moral outrage from Bohacek or many others now speaking up about Trump’s latest insult. Why? Because those weren’t their children, and those children’s suffering didn’t personally inconvenience them.
I found another example: Fred Trump III, the president’s nephew, also spoke out about the slur. He has a disabled son and lobbied actively for disability funding during the first Trump administration. That’s his single issue. But family separation? Children in cages? Apparently not on his radar, despite having direct access to power when those policies were being implemented.
The pattern is clear: these aren’t principled stands against cruelty or injustice. They’re personal grievances dressed up as moral positions. And in Bohacek’s case, he’s willing to hold redistricting hostage—treating his vote, which belongs to his constituents, as a personal poker chip to settle scores based on wounded feelings rather than policy merits. If your moral alarm only activates when the target looks like your own child, that’s not empathy—that’s narcissistic perimeter-setting.
What This Reveals About Human Reasoning
When I run these sanity checks with AI, I’m not looking for moral instruction from machines. I’m using them as tools to test whether my confusion about illogical human communication is warranted. And increasingly, I’m finding that yes, some elected officials really are this incoherent, this narrow in their thinking, this incapable of reasoning beyond personal experience.
A DePaul education apparently didn’t teach Senator Bohacek basic logical argumentation: state your claim, provide relevant evidence, explain the connection. His post fails all three. And if this is the quality of reasoning he brings to legislative decisions—conflating unrelated issues, substituting autobiography for argument, explicitly using legislative votes as personal weapons—that’s alarming regardless of his moral limitations. You cannot make sound policy decisions if you cannot think clearly, and you absolutely cannot serve the public if you’re willing to sabotage democratic processes to discipline someone who hurt your feelings.
The irony is rich: language models trained purely on text can demonstrate more consistent moral reasoning than humans with full sensory experience and decades of life. We don’t need firsthand experience of every injustice to recognize it as wrong—that’s what literature, history, journalism, and ethical education are supposed to cultivate. The capacity for empathy based on secondhand knowledge is fundamental to human moral reasoning. It’s what separates us from primitive organisms that can only respond to direct stimuli and lack the benefit of collective memory.
Yet here we have an elected official operating exactly like those primitive organisms—unable to render judgment except through firsthand experience, unable to extend moral concern beyond his immediate family, unable to construct logical arguments for positions he claims to hold, and willing to weaponize legislative power for personal vendettas.
If an elected official demonstrates less moral reasoning, less perspective, and less logical consistency than a statistical model, the issue isn’t that AI is too capable; it’s that some humans in power aren’t doing the minimum required of their role.
The Broader Implications
This matters beyond one incoherent Facebook post. This is a person making decisions about healthcare, education funding, disaster relief, immigration policy, and countless other issues affecting millions of lives. If he can only understand suffering he’s personally experienced, he’s fundamentally unqualified to represent diverse constituencies facing challenges the representative will never encounter. And if he’s willing to vote against his constituents’ interests to punish someone for personal slights, he’s actively dangerous in office.
It’s also deeply troubling from any religious or philosophical standpoint that should inform public service. From the little I know about Christianity, Christian teaching emphasizes care for strangers and enemies, expanding moral concern beyond your tribe—the entire point of parables like the Good Samaritan. “Love your neighbor as yourself” doesn’t mean “love people who remind you of your child.” Yet this narrow self-interest combined with vindictive score-settling is what passes for principled advocacy.
When I encounter posts like Bohacek’s that make me do a double-take, I’m grateful for AI tools that help me confirm I’m not losing my grasp on logic. The whiplash I felt reading that redistricting-to-daughter-to-Trump hostage-taking wasn’t a language barrier—it was a recognition that the post made no sense because the thinking behind it made none, and the intent behind it was to privatize public office for personal grievances.
Sometimes the most valuable thing AI can do is help you see that your confusion is warranted, that some humans really are thinking this poorly, and that maybe we should expect more from decision-makers than moral reasoning inferior to a language model’s pattern-matching—and certainly better than holding democratic processes hostage to discipline party leaders who’ve hurt their feelings.
[This post was developed through conversations with Claude Sonnet 4.5, ChatGPT-5.1, and Gemini 3 Pro and drafted with assistance from Sonnet 4.5 and GPT-5.1.]
ChatGPT-5.1
Prompt: I found this post on Instagram:
Republican state Sen. Mike Bohacek is speaking out against President Donald Trump after the president referred to Minnesota Gov. Tim Walz as “seriously r——-.” On Friday, Nov. 28, Bohacek posted a statement on Facebook denouncing Trump’s use of the ableist slur, which appeared in a Thanksgiving message on Truth Social the previous evening attacking Walz, 61, and his record on immigration. “Many of you have asked my position on redistricting,” Bohacek wrote. “I have been an unapologetic advocate for people with intellectual disabilities since the birth of my second daughter.”
Curious about your take on this?
Gemini 3 Pro
Prompt: To me, this person seems like a non-reasoning (primitive?) organism which can only render judgment based on firsthand experience. You don’t have to agree with me, but I think you’re giving him too much credit. The wording is proof that this person is incapable of the minimum level of reasoning required of a public servant:
“I have been an unapologetic advocate for people with intellectual disabilities since the birth of my second daughter.”
He’s only been advocating for people with certain disabilities (not all, because what have those other people done to deserve his support?) since the birth of the second daughter with that specific challenge.
MUCH too narrow. This guy wouldn’t even be a good custodian for his own family, since he hasn’t personally experienced child birth, Down syndrome, or gender inequity.
Oh, and let’s not forget that this person’s extremely self-centered advocacy has nothing to do with his vote on the redistricting issue. That post had me scratching my head because I experienced a whiplash (What am I missing here? Those two things have nothing to do with each other?!)
And isn’t the same kind of “I’m ok, because it’s just the others” complacency that people in Nazi Germany realized got them into that whole mess? Exactly the wrong mindset that you would never want in a public servant.
Flash and you are the outliers on this in thinking it was effective political communication. I’d hate to see you being right in giving this person credit for understanding his electorate (that they’ll see politicians taking principled stances without any personal stakes as inauthentic). I’d like to believe that US voters will be able to see that this person has not outgrown a young kid’s egocentric view of the world and is unable to think through consequences of policy decisions he is regularly called upon to make (”entitlements,” health insurance, disaster relief, school lunches [which he’ll probably say he’s against because that’d keep you from pulling yourself up by your boot straps :D]). So he’s unqualified to represent voters. I feel sorry for his kids, for whom he’s setting exactly the wrong example. Also deeply wrong (from the little I know about Christian values, since I’m an agnostic) for any person of faith to be so self-centered. I thought the teachings of Jesus, not his personal situation, should be informing his life?
Claude Sonnet 4.5
Prompt: Whoa, the actual post was even sillier than we all thought. This person has no business representing the people of Indiana!
“Many of you have asked my position on redistricting. I have been an unapologetic advocate for people with intellectual disabilities since the birth of my second daughter. Those of you that don’t know me or my family might not know that my daughter has Down Syndrome. This is not the first time our president has used these insulting and derogatory references and his choices of words have consequences. I will be voting NO on redistricting, perhaps he can use the next 10 months to convince voters that his policies and behavior deserve a congressional majority.”
