Artificial intelligence has quietly colonized every stage of the American medical billing pipeline. AI scribes now listen to doctor-patient conversations and translate them into billing codes — codes that, according to Blue Cross Blue Shield research, systematically inflate charges. AI chatbots intercept patients calling about their bills. Machine learning models predict how much a patient can be squeezed to pay. And AI-driven companies help debt buyers price portfolios of unpaid medical bills. The technology is not incidental to the system. It is becoming the system. The timing is not accidental. Republican-led health cuts through the One Big Beautiful Bill Act slashed roughly $1 trillion from federal healthcare spending. Paired with the refusal to extend Obamacare marketplace subsidies, the Kaiser Family Foundation estimates about 17 million people will lose insurance coverage through 2034. The Urban Institute projects $278 billion in uncompensated care over the same period — bills that hospitals know won't be paid but will now be pursued with unprecedented technological sophistication. Cedar, an AI billing platform backed by Andreessen Horowitz and JP Morgan, exemplifies the new model. The company uses hospital data to train machine learning models predicting a patient's 'propensity to pay,' then matches patients with 'AI-driven discounts' — personalized pricing for medical debt that mirrors the surveillance pricing the FTC has flagged in consumer markets. Cedar deploys an AI chatbot named Kora to handle billing inquiries, cutting hospital labor costs while maintaining collection pressure. The coding side is equally consequential. AI scribe companies listen to exam-room conversations and suggest enhanced billing codes. One such company, Ambience Healthcare (also Andreessen Horowitz-backed), claimed its software generated an additional $13,000 in revenue per clinician per year through 'coding accuracy.' A second BCBS analysis estimated AI coding added $1 billion in healthcare costs in a single year. The American Hospital Association disputes that this constitutes inaccurate coding, but the directional incentive is unmistakable: AI optimizes for revenue capture. Insurers are deploying the same technology in reverse. AI algorithms are used to delay and deny care through opaque prior authorization systems. Robert F. Kennedy Jr.'s health department recently launched an AI-powered prior authorization program for Medicare, institutionalizing a tactic previously confined to private insurers. Patients now face AI on both sides — one system maximizing what they're billed, another minimizing what's covered. More than one in four US adults reported owing medical debt in 2025, according to a joint analysis by the Urban Institute and the Robert Wood Johnson Foundation. The proportion held for insured patients, undermining the assumption that coverage protects against debt. As deductibles, co-pays, and out-of-pocket costs rise, insurance increasingly functions as a fee for partial access rather than financial protection. JP Morgan has celebrated AI in coding, billing, and collections as 'early wins.' The framing reveals who the technology serves. Every efficiency AI creates in this ecosystem — faster coding, cheaper collection calls, more precise pricing of patient desperation — flows upward to hospital revenue departments, venture-backed billing companies, and debt buyers. The patient receives a chatbot.
