Artificial intelligence has become a flashpoint in political discourse, with many on the left viewing it with suspicion—as a tool of corporate power that exacerbates inequality, displaces jobs, and entrenches surveillance. Yet a growing minority of progressives argue that AI can be a force for good, particularly when wielded to expose flaws in mainstream media narratives and advance working-class agendas. One of the most compelling examples of this comes from Matt Bruenig, a socialist lawyer and policy analyst who has used AI to challenge a prominent Wall Street Journal article about young Americans moving back in with their parents.

Bruenig’s work, detailed in a lengthy YouTube video, demonstrates how AI can serve as a low-cost, high-efficiency research assistant for anyone seeking to fact-check corporate media. The video centers on a Wall Street Journal story that highlighted a startling statistic: 49% of adults under age 30 lived with a parent in 2025, up 12 percentage points from 2019. The article framed this shift not as a crisis, but as a pragmatic lifestyle choice—a way for young people to save money or regroup after a breakup. The implication was that attitudes had changed, making multigenerational living a normal, even desirable arrangement.

Bruenig, however, smelled something amiss. He suspected that the dramatic 12-point jump might not reflect a real cultural shift but rather a flaw in the underlying survey data. To test his hypothesis, he turned to Claude Code, an AI-assisted coding tool, his usual terminal emulator Ghostty, and the tmux manager. He had previously loaded the raw microdata from the Federal Reserve’s Survey of Household Economics and Decisionmaking (SHED) into folders managed by the file manager Yazi. With a simple verbal prompt—essentially asking Claude to investigate whether the statistic was fishy—the AI went into what he called “bloodhound mode.”

The Mechanics of the Analysis

Claude Code rapidly parsed the SHED data and identified what Bruenig describes as a “surveying artifact.” The issue lay in how the survey’s age brackets were constructed. In one year, the 18–29 age group contained disproportionately few teenagers—the demographic most likely to live with parents. In another year, the age distribution was more balanced. This imbalance, when plotted over time, created the illusion of a sudden 12-percentage-point jump. In reality, the underlying trend was far more modest: a genuine increase of roughly 3 percentage points, consistent with the ongoing affordability crisis in housing, not a sudden attitudinal revolution.

“What the AI discovered is that the Journal’s headline stat is basically noise,” Bruenig says in the video. “It’s a classic example of how bad sampling can lead even the best journalists astray.” The analysis underscores a broader point: AI can augment human skepticism, rapidly performing data checks that would take a lone researcher hours or days. For leftists who lack the resources of major newsrooms, such tools could level the playing field.

Bruenig’s video is not for the faint of heart. Running 51 minutes, it dives deep into statistical weighting, survey methodology, and the quirks of SHED’s questionnaire design. But for those willing to engage, it offers a transparent, reproducible demonstration of AI-assisted fact-checking. He walks viewers through each step—how he instructed Claude to visualize age distributions, how the AI flagged anomalies, and how he verified the results using standard statistical tests.

Why This Matters for the Left

The left’s reluctance to embrace AI stems from understandable concerns. Tech giants dominate the development of these tools, and their underlying models often reflect biases baked into training data. There is also the risk that AI will accelerate job displacement in sectors like customer service, warehousing, and content creation. Yet Bruenig argues that abstaining from AI is a strategic mistake. “The right is already using these tools to shape narratives, build databases, and organize,” he told Transformer’s Dan Kagan-Kans earlier this year. “If the left refuses to engage, we’ll be left behind in the information war.”

His work with Claude Code provides a concrete example of how AI can serve leftist ends: debunking harmful myths, exposing statistical sleight of hand, and reinforcing evidence-based arguments for policies like rent control or public housing. Consider the Wall Street Journal’s framing: by suggesting that young people are choosing to live with parents, the article tacitly downplays the severity of the housing crisis. If the real increase is only 3%, then the story becomes less about cultural change and more about economic desperation—a narrative that supports calls for increased housing subsidies and tenant protections.

“This is exactly the kind of work that grassroots organizers and progressive think tanks should be doing,” says Bruenig. “But they don’t have the staff or the budget for extensive data analysis. AI can fill that gap.” Indeed, the entire analysis was conducted on a standard laptop, using free or low-cost tools. The code and data are publicly available, allowing anyone to replicate the findings. This transparency aligns with leftist values of openness and community accountability.

Broader Implications for Media Criticism

The Wall Street Journal article is far from an isolated case. Mainstream media frequently relies on survey data that may be flawed, either due to sampling errors or intentional spin. AI tools like Claude Code, paired with statistical literacy, could become a powerful weapon for media watchdogs. Imagine a future where progressive fact-checkers routinely deploy AI to scrutinize claims in major news outlets, from unemployment figures to crime statistics to polling numbers.

Of course, AI is not a panacea. It can amplify biases as easily as it corrects them. Leftists must remain vigilant about the datasets and algorithms they use. Bruenig’s approach—publicly sharing his methodology and inviting criticism—offers a model for responsible use. He also emphasizes that AI should not replace human judgment. “The AI caught the artifact, but I had to interpret it,” he notes. “It’s a tool, not a magician.”

The housing statistic exposure is particularly timely. With young people in many countries facing skyrocketing rents and stagnant wages, the narrative that they are “choosing” to live with parents is politically convenient for those who oppose government intervention. By debunking this narrative, Bruenig’s AI-assisted research provides ammunition for activists pushing for more aggressive housing policies. It also serves as a reminder that numbers don’t speak for themselves—they require skeptics with the tools to interrogate them.

A Practical Demonstration for Skeptics

For leftists still on the fence about AI, Bruenig’s video offers a practical, low-stakes entry point. It does not require anyone to embrace generative AI for creative work or to trust chatbots with sensitive organizing data. Instead, it showcases a narrow, technically constrained use case: using an AI coding assistant to automate statistical analysis. This is hardly the stuff of dystopian science fiction. It is more akin to using a calculator or a spreadsheet—a productivity enhancer that frees up mental bandwidth for more strategic thinking.

Moreover, the tools are becoming easier to use. Claude Code, like competing products from GitHub Copilot and OpenAI’s Codex, can be operated via natural language instructions. You don’t need a computer science degree to prompt it to “check if the age distribution in the survey is consistent over time.” The AI handles the messy coding details, while the human focuses on the question’s framing and the logic of the analysis. This lowers the barrier for non-technical activists to engage with data.

Bruenig has been a vocal proponent of this approach. In his earlier writing on the topic, he argued that the left’s traditional suspicion of technology is a luxury it can no longer afford. “We need every advantage we can get,” he wrote. “If the right is using AI to draft legislation and target voters, we’d better learn how to use it too—or we’ll be fighting with one hand tied behind our backs.” His video provides a blueprint for how to start.

As housing debates heat up in cities from New York to San Francisco, the ability to quickly dissect mainstream media claims will be crucial. The Wall Street Journal piece is just one skirmish in a larger war over how we understand the housing crisis. Bruenig’s contribution—with AI as his ally—shows that the left need not cede the battlefield of data and evidence to the right.

Whether you agree with him politically or not, the video is a masterclass in citizen fact-checking. It raises important questions about who gets to tell the story of America’s housing market and how robust those stories really are. And it suggests that AI, for all its flaws, can be a magnifying glass for truth—if wielded by someone who knows exactly what they’re looking for.

Source: Gizmodo News