AI Complaint Systems May Unfairly Judge You By Writing Style

Imagine two people file the same complaint about a bank fee. One writes in formal, polished English; the other uses casual language with a few grammar mistakes. According to a new study published on November 6, 2026, on ArXiv, artificial intelligence (AI) systems used to decide these cases may give different outcomes—just because of how each person writes. That means your education, age, or background could influence whether your complaint is taken seriously or dismissed, even if the facts are identical.

What Happened

The study, titled "Same Facts, Different Remedy: Linguistic Profiling and Arbitrariness in LLM and System One Decisions," looked at how large language models (LLMs) like GPT-4o-mini and GPT-5-mini handle real consumer complaints. Researchers took actual complaints from the US financial sector and rewrote them in different language styles, keeping every fact the same. The only thing that changed was the style—formal, casual, or reflecting different education or socioeconomic backgrounds.

When these rewritten complaints were run through AI systems, the outcomes varied. In some cases, the AI sided with the complainant when the writing was polished, but denied the same complaint when it was written in a less formal or less "standard" style. This shows that AI isn't just looking at the facts—it's also making judgments based on how people write. The study raises red flags about fairness and bias, especially as more companies turn to AI to handle customer service and complaints.

This isn't just a theoretical problem. Major banks and insurance companies in the US have already started using AI to sort and decide on customer complaints, hoping to save money and speed up decisions. But if these systems are biased, regular people could be losing out on refunds, corrections, or fair treatment simply because of their writing style.

How This Affects Everyday People

For workers, parents, students, and anyone who has ever filed a complaint or asked for help from a company, this research hits close to home. Many people—especially those who didn't go to college, speak English as a second language, or just don't write in a formal style—could be at a disadvantage. Your ability to get a refund, fix a billing error, or resolve a problem might depend on how "professional" your complaint sounds to an AI, not just on whether you're right.

Take the example of a parent disputing a medical bill. If they write their complaint in plain language with a few spelling mistakes, the AI might be less likely to side with them than if the same complaint was written in perfect English. The same goes for older adults who might use different phrasing, or young people who write more casually. This kind of bias can hit hardest in communities that already face barriers—like lower-income families, recent immigrants, or people with disabilities.

Students and young workers, too, might find themselves at a disadvantage when dealing with banks, landlords, or service providers. If AI systems are making decisions based on writing style, those still learning English or who haven't had much practice with formal writing could be unfairly penalized. This could mean lost money, denied services, or more stress for people who are already struggling.

Real-World Examples

• A gig worker disputes a late fee with their bank. Their informal email is dismissed by the AI, while a more polished version of the same complaint gets a refund.
• A recent immigrant tries to resolve a utility billing error. The AI doesn't recognize the urgency in their message due to non-standard phrasing, so the complaint is deprioritized.
• An elderly customer writes in the style of their generation. The AI misinterprets their tone and denies their request, even though the facts are clear.

These aren't just hypotheticals—they're the kind of cases the ArXiv study found in its audit of AI systems.

The Bigger Picture

This isn't the first time AI bias has made headlines. In 2023, the Consumer Financial Protection Bureau (CFPB) warned banks about using AI in lending decisions, citing risks of discrimination under the Equal Credit Opportunity Act. In 2024, the European Union passed the AI Act, requiring companies to test AI for bias and explain high-stakes decisions. But as this new research shows, even the latest AI models can still be swayed by factors like writing style—something that isn't always covered by current regulations.

AI is now being used not just in customer service, but in hiring, housing, and even the justice system. If these systems are biased by language style, the risks go far beyond consumer complaints. It could mean unfair job rejections, denied housing applications, or unequal treatment in court. As more companies automate decisions, the potential for "linguistic profiling"—judging people by how they write—becomes a real civil rights issue.

For workers worried about AI taking over their jobs, this is another reminder that AI isn't just replacing human labor—it can also bring new forms of unfairness. And for anyone tracking how AI is changing the workplace, it's a sign that technology isn't always neutral, even when it claims to be. For more on how AI is displacing jobs, see our AI jobs explainer.

What You Can Do

• Be clear and detailed in your complaints. If you need to file a complaint, try to include all relevant facts and use clear, direct language. If possible, have someone review your message before you send it.
• Ask for human review. If your complaint is denied or ignored, specifically request that a human look at your case. Many companies still offer this option, even if it's not advertised.
• Know your rights. In the US, you can file complaints with agencies like the Consumer Financial Protection Bureau (CFPB) if you believe you've been treated unfairly by a financial company. In the EU, the AI Act gives you the right to an explanation for automated decisions.
• Support fair AI rules. Contact your representatives to ask for stronger laws that require companies to test AI for bias and provide human oversight. The more people speak up, the more likely lawmakers are to act.
• Share your experience. If you think you've been treated unfairly by an AI system, tell your story online or to consumer advocacy groups. Real examples help push for change.

The Bottom Line

AI is supposed to make life easier, but when it comes to handling complaints, it can introduce new forms of unfairness. As this latest research shows, the way you write—not just what you say—can sway AI decisions that affect your money, your services, and your rights. Until stronger rules are in place, everyday people need to stay alert, speak up, and demand fairness from the technology that’s supposed to serve us all.