TL;DR — Key Takeaways
– AI-driven hospital billing adds nearly $1 billion in costs: A Blue Cross Blue Shield Association report found that increased use of AI-assisted medical coding added $942 million in expenses for insurers between 2024 and 2025 compared with 2023 levels.
– Secondary diagnoses drive higher reimbursements: Of the additional costs, $653 million came from increased documentation of secondary medical conditions, including anemia and low sodium.
– Higher billing does not necessarily reflect sicker patients: Despite increases in complex diagnoses, corresponding treatment rates remained largely unchanged, raising questions about whether increased coding intensity reflects actual changes in patient health.
Widespread adoption of artificial intelligence (AI) in healthcare settings has added nearly $1 billion in extra expenses for health insurers over the past two years, according to a report released Thursday by the Blue Cross Blue Shield Association (BCBSA).
The study, which analyzed inpatient hospital billing across BCBSA’s network of 31 independent insurers covering more than 100 million Americans, found that providers are increasingly using AI algorithms to inflate patient complexity and claim higher reimbursements.
Between 2024 and 2025, healthcare facilities increasingly leveraged AI tools such as ambient clinical scribes and automated record scanners to flag coexisting or secondary conditions. The shift toward more intense care documentation added $942 million in total costs compared to 2023 levels, with $653 million driven directly by secondary condition billing.
By adding diagnoses such as anemia or low sodium to existing records, hospitals increased payments by an average of nearly $12,000 per case. For patients undergoing major bowel surgeries, reported secondary conditions like partial intestinal blockages and acid overload surged by 55% and 33%, respectively, between early 2023 and late 2025.
However, clinical data indicates these higher billing claims do not reflect sicker patients.
“If patients are truly sicker, we’d expect to see more treatment,” said Luke Chalker, senior vice president of product and data science at BCBSA. “The disconnect between diagnoses and treatment suggests that AI is identifying more billable conditions, not sicker patients.”
Dr. Razia Hashmi, BCBSA’s vice president of clinical affairs, noted that despite the spike in complex diagnoses among bowel surgery patients, actual care rates remained flat. For instance, new diagnoses of anemia were not accompanied by an increase in blood transfusions, the standard treatment for the condition.
The report highlights a growing coding war between providers using AI to maximize revenue and insurers using similar automated tools to flag aggressive billing or deny care authorizations.
Major insurers, including Centene, have criticized the practice as inappropriate, while hospital systems like Michigan-based McLaren Health Care openly acknowledge using AI to boost revenues by $1 million monthly through enhanced documentation.
Industry experts warn that this automated friction — where low-cost AI agents continuously challenge each other over reimbursements — will ultimately push administrative overhead onto employers and consumers, with projections pointing to double-digit health cost increases next year.
