Patients are turning to AI chatbots to contest soaring medical bills, a trend that has caught hospitals’ attention. Jackie Davalos and partner Walter Kerr fed the Anthropic bot Claude with a $22,604 emergency‑room invoice from George Washington University Hospital. Claude returned a list of legal angles, prompting Kerr to write a demand letter that ultimately led the hospital to erase the debt.

The American Hospital Association recently warned members that more patients are using generative AI to dispute charges, underscoring a broader friction between providers and insurers that have long deployed AI to maximize reimbursements. Critics argue bots can misinterpret statutes and lack HIPAA safeguards, leaving inexperienced users vulnerable. OpenAI claims newer ChatGPT models now request missing details to curb such errors.

Success stories like Davalos’s are offset by cases such as Atlanta‑based Michelle Maziar, whose ChatGPT query missed her employer’s self‑insured status and offered dead‑end steps. Legal aid experts view AI as a double‑edged sword: it can translate jargon and spot billing errors, yet without human review it may steer patients away from effective remedies. The market feels pressure to make health‑care bots accurate and compliant.