A new survey reveals that one in four insured Americans have asked an AI chatbot for medical advice because they couldn’t afford to see a doctor, and when they do, ChatGPT is their overwhelming first choice. The findings, from a Pollfish survey of 1,250 U.S. adults conducted for Insuranceopedia, highlight both the growing role of consumer AI in healthcare navigation and the risks of substituting chatbots for clinical care.

ChatGPT dominates—but Google Gemini is the only real challenger

When consumers turn to AI for help understanding their health insurance or medical bills, platform choices are far from evenly distributed. Among those who expressed a preference, 28% named ChatGPT, while 21.3% chose Google Gemini. Together, the two leaders account for nearly half of the market, leaving every other assistant in single digits.

Microsoft Copilot—deeply embedded in Windows, Microsoft 365, and Edge—managed just 6.4%. Meta AI followed at 4%, Claude at 2.3%, Grok at 1.6%, and Perplexity and other tools each below 1%. The lopsided results, first reported by TechTarget, show that general consumer mindshare, not platform integration, drives which chatbot patients trust with sensitive health questions.

ChatGPT’s lead mirrors its broader cultural position: for many, “ChatGPT” has become synonymous with “AI assistant,” much like “Google it” became shorthand for web search. Gemini benefits from Google’s search dominance and Android reach. Copilot, meanwhile, is often perceived as a workplace tool rather than a personal companion—a branding gap that could impede its adoption in consumer health.

The real headline: people are using AI because they can’t afford care

Beyond platform rankings, the survey contains a far more urgent finding. A full 25% of respondents said they had asked an AI chatbot a medical question specifically because they could not afford to see a doctor. That figure—one in four insured adults—suggests that cost-driven reliance on AI is not a fringe behavior.

The danger deepens downstream. Among those who turned to AI because of cost, 40% reported that they later skipped, delayed, or sought treatment based solely on the chatbot’s guidance. That means millions of people may be making healthcare decisions with no human clinical oversight.

Age further segments the risk. Younger adults are far more likely to lean on AI out of financial necessity:

  • Ages 18–24: 53.1%
  • Ages 25–34: 41.2%
  • Ages 35–44: 38.2%
  • Ages 45–54: 28.4%
  • Ages 55–64: 16.1%
  • 65 and older: 6.8%

This pattern isn’t merely a low-income phenomenon. Many respondents in middle-income brackets reported similar pressures, reflecting high deductibles, copays, and unpredictable out-of-pocket costs that make a clinic visit financially daunting even for the insured.

Why chatbots feel like a better option than insurance company hotlines

Respondents offered clear reasons for their AI preference. A third valued 24/7 availability. Another 30% said the chatbot explained medical or insurance concepts better than a human payer representative. And 28% noted that reaching an AI was easier than contacting their insurer.

Those frustrations are real. Insurers often require phone-tree navigation, long hold times, and business-hours-only access. AI tools, in contrast, respond instantly in plain English, let users restate questions, and never rush them off a call. For patients confused by an explanation of benefits or a claim denial, a chatbot can translate jargon, summarize a document, and generate a list of questions to ask a human—all without the friction of traditional channels.

But the same ease of use creates a trap. A well-written, confident-sounding AI answer can feel authoritative even when it’s wrong, incomplete, or inapplicable to a specific plan. Insurance policies vary by employer, state, network, and plan type; a generic explanation of “out-of-network deductible” may not reflect the user’s actual coverage.

What this means for Windows users and the Microsoft ecosystem

Copilot’s modest 6.4% share is a reality check for Microsoft. The assistant is a central pillar of the Windows 11 experience, integrated directly into the taskbar and bundled with Microsoft 365. Yet when consumers need help with something as personal as a medical bill, they tend to reach for a tool they associate with conversation, not spreadsheets.

Microsoft’s healthcare opportunity isn’t to win a direct consumer chatbot shootout. It lies in making Copilot genuinely useful inside the systems patients already use—secure patient portals, billing websites, appointment scheduling tools—under strict governance. When connected to verified, plan-specific data with patient consent, a healthcare-aware Copilot could explain where to check a deductible balance and offer a seamless handoff to a human representative. That’s a far more valuable role than serving up generic definitions.

For everyday Windows users, the takeaway is practical: if you choose to use Copilot, ChatGPT, or any chatbot to interpret a medical document, treat the response as a starting point—never as a final answer.

Privacy risks are growing as patients upload sensitive records

Patients are increasingly comfortable feeding sensitive information into AI chatbots. The survey didn’t measure exactly what people upload, but typical use includes explanation of benefits forms, bills, lab results, medication lists, physician notes, and even full insurance member numbers.

Any such data shared with a consumer chatbot could be retained, used for training, or exposed in a breach unless the platform offers specific, verifiable healthcare privacy protections. Users should be aware that not all “HIPAA-compliant” claims apply to ordinary chat sessions, and that privacy settings vary by feature, account tier, and connected add-ons.

At minimum:
- Remove names, member IDs, dates of birth, and account numbers before uploading.
- Avoid sharing documents that contain multiple people’s protected information.
- Use temporary or expiring chats when available.
- Check whether the service stores chat history and who can access it.

Practical steps for using AI safely with health insurance

AI can genuinely reduce the overwhelm of dealing with healthcare bureaucracy—but only when used as a guide, not a decision-maker. The most reliable workflow:

  1. Use AI to translate and prepare. Ask it to define terms, summarize a bill, or list questions to ask your insurer. Treat its output like a smart friend’s suggestion, not a policy determination.
  2. Verify against primary sources. Go back to your plan’s portal, member handbook, or insurer’s customer service for confirmation. A chatbot cannot guarantee coverage or process an appeal.
  3. Keep humans in the loop. The survey found 35% of people would never use AI for health information. That’s a reminder that phone lines, patient portals, and in-person help are still essential. Insurer call centers and provider billing offices remain the final authority.
  4. Escalate when necessary. If a claim is denied, medication is unaffordable, or a treatment is in question, file a formal appeal and seek assistance from consumer-health nonprofits, state insurance departments, or employer benefits advocates.
  5. Don’t substitute AI for a doctor. This is the hardest rule when cost is a barrier, but a chatbot cannot perform a physical exam, check vitals, or assess nuance. If a symptom could be serious—chest pain, sudden vision changes, difficulty breathing—seek emergency care, not a prompt.

Outlook: what to watch next

The survey captures a moment in which AI is filling a gap created by healthcare affordability and accessibility problems. Expect regulatory attention to intensify, especially around AI tools that blur the line between information and medical advice. The Federal Trade Commission and state attorneys general have already signaled interest in AI-related consumer harm.

Microsoft, meanwhile, will need to decide whether to compete head-on for consumer health mindshare or to double down on governed, enterprise-grade healthcare deployments where Copilot’s identity-management and compliance strengths shine. Either path requires clearer guidance for users on privacy, accuracy, and when to seek human help.

For patients, the best future isn’t one without AI—it’s one where a chatbot helps you understand your paperwork, and a real person helps you understand your health.