Will AI fix prior authorization—or make it worse?
Artificial intelligence is increasingly being deployed to manage "prior authorisation," the process by which health insurers pre-approve medical treatments, but experts are divided over whether it will ease or worsen a system already widely criticised by patients and doctors. While AI could theoretically speed up approvals for clearly valid claims and reduce harmful delays, critics warn it may instead be used to automate and increase wrongful denials of necessary care, prompting calls for greater transparency and accountability from insurers.
A 2025 American Medical Association survey found 61% of doctors feared AI would worsen denials of treatments they considered necessary, with the AMA pushing for insurers to give detailed clinical justifications and disclose more about their algorithms. In Medicare Advantage, which covers roughly 55% of eligible seniors and disabled Americans, insurers issue millions of denials annually, and a June government report found some plans even reject requests for skilled nursing and rehabilitation care. A Commonwealth Fund survey found around one in five privately insured working-age adults reported a denial affecting themselves or a family member in 2025, with 41% saying it delayed care and over a quarter saying their condition worsened as a result. The Trump administration is piloting an AI-driven cost-reduction programme in six states, while a 2024 Biden-era rule set new turnaround limits, requiring decisions within 72 hours for urgent requests and seven days for non-urgent ones.
In short
- AI is being tested to speed up or manage medical prior authorisation approvals
- 61% of doctors fear AI will increase wrongful care denials
- One in five insured US adults reported a 2025 care denial affecting their family
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Proponents argue that AI can standardise and accelerate prior authorisation decisions for straightforward cases, reducing administrative delays that demonstrably harm patients with clearly necessary care. By applying consistent clinical criteria more uniformly than human reviewers, AI could eliminate subjective denials whilst also helping identify genuinely unnecessary or fraudulent claims, ultimately improving system efficiency and patient outcomes when properly designed with clinical input and transparency safeguards.
The case against
Critics contend that deploying AI in this context risks automating and scaling up wrongful denials rather than improving approvals, given insurers' financial incentive to deny claims and the opacity of algorithmic decision-making. The evidence of widespread denials already affecting patient outcomes, coupled with the 61% of doctors fearing AI will worsen this problem, suggests the technology may simply enable faster, harder-to-challenge rejections of necessary care unless insurers face rigorous transparency requirements and genuine clinical accountability.
Explained simply
Prior authorisation is the system where health insurers must formally approve medical treatments before patients can receive them. It exists to help control costs, but is already widely criticised by patients and doctors because denials can delay or prevent necessary care.
Insurance companies issue millions of denials each year, with many for treatments doctors consider essential. Surveys found that about one in five working-age adults reported a denial affecting them or a family member in 2025, and roughly four in ten said it delayed their care, whilst over a quarter said their health worsened.
Healthcare providers are increasingly using artificial intelligence to process prior authorisation decisions, but experts are divided on whether this will help or harm patients. Supporters argue it could speed up approvals for clearly valid cases and reduce harmful delays, whereas critics worry it will simply automate and accelerate denials, and they are calling for greater transparency from insurers about how their algorithms work.
AI Government Politics Technology