AI Used to Collect Growing US Medical Debt
Left 50%
Center 0%
Right 50%
1 left · 0 center · 1 right
What happened
A hospital patient arranged medical care as a self-pay patient and paid the bill in full. The patient later received $27,000 in additional bills for those same services. On Oct. 6, the hospital said an AI-enabled process within its Epic electronic health records system had identified that the patient had insurance. That process billed the insurance anyway, despite the completed self-pay arrangement.
Sourcing check · Fully on the record · 2 articles
1 of 1 article cites a named source for the story. It was on the record from the first report.
What these terms mean and how they're counted
- On the record · names a source 1 of 1
- On the record means the article cites a source by name for the story's main claim: a person, or an institution speaking officially, such as a company statement or a named agency's spokesperson. The opposite is an unnamed source. The count is how many of the counted articles cite at least one such source; the names are everyone named across them, each once however the articles spell it. Here: April Kuehnhoff
- Primary evidence · cites a document or data 1 of 1
- Primary evidence means the article cites a record you could check yourself: a court filing or ruling, a dataset or report, a recording or transcript, an official statement or press release, market data. A quote is not a document. The count is how many articles cite one. Here: report
- Credits another outlet 0 of 1
- The article says the story comes from another outlet's reporting (“first reported by Reuters”). Those articles relay the story; they are not independent confirmation. Here: Each article presents its own reporting.
- Unnamed sources 0 of 1
- The main claim is attributed to people who are not identified: “people familiar with the matter”, “officials said”. The count is how many articles do this.
- How it developed
- When the first report landed and whether it named anyone; when the first named source and the first document appeared. Here: First report Oct 12, 12:00 UTC by The Guardian, naming its source. First document in the first report, from The Guardian (report).
- Not counted 1
- Opinion, analysis and video pieces, and articles too short to assess. Every count above is out of the counted articles. An article republished under a new headline is one article.
This is about who is cited, not whether the story is true. Each source below says what it names. How we count
Omitted — what each side leaves out
Unpacked
Left-leaning coverage never reports the $27,000 overbill that is the entire right-leaning story. The Guardian instead describes AI that increasingly "listens, codes, bills and collects" as medical debt grows under Republican health cuts. Right-leaning coverage in the Washington Examiner is a first-person account of an Epic process that billed insurance after the writer had already paid as a self-pay patient. That gap decides whether the reader sees a collections industry expanding or a settled bill being reopened by software.
Left-leaning coverage leads with medical debt growing as Republican-led health cuts take effect. Right-leaning coverage leads with the writer having paid, then facing new bills for the same care.
Were the extra charges reversed, and how many other self-pay patients did Epic bill the same way?
Bottom line
The Washington Examiner’s $27,000 Epic rebill after a paid self-pay visit does not appear in the Guardian. Readers of only left-leaning coverage would not know a settled hospital bill can be reopened by AI.
The receipts — 3 facts one side's coverage lacks
- A patient received $27,000 in bills for services already paid
- The hospital explained the extra billing in a conversation on Oct. 6
- The patient had arranged and paid for the services as a self-pay patient
The Left View
The Guardian frames hospital AI as an expanding coding and collections apparatus that will be used more as medical debt grows, not as a one-off billing glitch. It reports that more than one in four U.S. adults owed medical debt in 2025, including among people with insurance, and quotes National Consumer Law Center attorney April Kuehnhoff saying debt will grow among "both the insured and uninsured, because of how we're increasingly putting more costs onto individuals for those plans." It ties that outlook to Republican-led cuts: Donald Trump's HR1, the One Big Beautiful Bill Act, "cut roughly $1tn in federal spending from the healthcare industry," and with the refusal to extend Affordable Care Act marketplace subsidies is expected to leave about 17 million newly uninsured through 2034 and, per the Urban Institute, $278bn in uncompensated care by 2034, after which "cash-strapped hospitals" will turn to "increasingly high-tech collectors." It describes firms such as Cedar training models on a patient's "propensity to pay" and matching patients with "AI-driven discounts," plus a chatbot, Kora, to speak with patients about bills and contain labor costs. On coding, it cites a Blue Cross Blue Shield Association white paper finding women in childbirth "increasingly being billed after severe blood loss during delivery even as the rate of blood transfusion rates remained steady," with researchers arguing AI was "boosting hospital billing"; a scribe company claimed an additional "$13,000 in revenue per clinician per year" through greater "coding accuracy," and a later BCBS report "claimed AI coding led to an additional $1bn in healthcare costs in a year." The American Hospital Association said those analyses "do not demonstrate that hospitals are coding patients inaccurately or that AI tools are driving inappropriate spending."
The Right View
The Washington Examiner, in a first-person account, presents the extra bills as an "'Epic' failure" in which hospital AI overrode a completed self-pay arrangement and billed insurance anyway. The writer says that explanation "should concern every American who has ever opened a medical bill." The piece treats the harm as an automated Epic process generating tens of thousands of dollars in charges for care already paid for, rather than as a dispute over coverage policy or collection strategy.
Our Take (balanced)
The left's strongest argument is that AI is being built to code more aggressively and collect more as uncompensated care rises: Cedar's "propensity to pay" models, a vendor's claimed additional "$13,000 in revenue per clinician per year" from "coding accuracy," and BCBS findings that childbirth cases were "increasingly being billed after severe blood loss" while transfusion rates "remained steady," set against Urban Institute estimates of $278bn in uncompensated care after HR1. The right's strongest argument is the self-pay episode itself—after payment in full, Epic's AI still identified insurance and produced $27,000 in further bills. The central unresolved tension is whether hospital AI is primarily an extraction engine scaling with coverage-policy shifts that push costs onto patients, or an automated billing layer that can ignore a finished payment arrangement and manufacture large erroneous charges.
2 sources
More stories
The week's bottom lines, in your inbox
One email a week: the five stories that mattered and what they actually mean. Free.