The AI arms race in healthcare: Insurers deploy bots to deny care, patients deploy bots to fight back

🌟 Editor's Note

Healthcare's AI battle lines just got drawn. UnitedHealth, Elevance, and Centene deploy AI across 500+ tasks while denying 19% of claims. Patients fight back with AI tools overturning 1,000+ denials. Johns Hopkins cuts $1.4M in costs, and Trump tests AI Medicare gatekeepers in 6 states starting January 2026. Welcome to healthcare's AI arms race where both sides automate, and patients are caught in the middle.

Inside the nation's largest health insurers: AI adoption is taking off like a rocket

UnitedHealth uses AI in 500 tasks, with CEO Andrew Witty revealing employees use it to write millions of lines of code and speed decisions on costly medicines. "Every use case is delivering at least double-digit percentage rates of efficiency," Witty said.

Elevance rolled out generative AI to 50,000 employees and hired 500 people exclusively for AI. Centene uses AI to manage contracts for 28 million people. However, all five major insurers declined to elaborate on AI use, creating a transparency crisis as denial rates climb to 19%.

Read the investigation: Health insurers' AI explosion →

AI vs. AI: Patients deploy bots to battle health insurers that deny care

As insurers weaponize AI, patients arm themselves with tools to fight denials. Counterforce Health's free AI assistant analyzes denials, reviews policies, and drafts appeals—overturning roughly 1,000 denials. Fight Paperwork generated 6,000+ appeals since May, reducing charges from $1,000 to $20 in one case.

The stakes: Insurers denied 19% of in-network claims in 2023, yet fewer than 1% of consumers appealed. Four states now ban AI-only decisions; a dozen+ states passed AI healthcare laws this year.

"We don't want to get on an AI-enabled treadmill that just speeds up," said Harvard Law's Carmel Shachar. "Two robots arguing over whether a patient should access care doesn't feel like a satisfying outcome."

Read the full story: Patients fighting back with AI →

UC Health study reveals surprising truths about how patients actually use healthcare chatbots

Matthew DeCamp, MD, Ph.D. at CU Anschutz studied UC Health's chatbot Livi, revealing patients use chatbots differently than healthcare systems expect. The JAMIA study shows surprising patterns in question types, trust levels, and willingness to act on AI recommendations.

"Understanding how consumers actually behave with these tools—not how we think they should behave—is critical for designing effective healthcare AI," DeCamp explains.

Read the coverage: How patients use healthcare chatbots →

Read the study: JAMIA research →

Trump administration to test AI Medicare gatekeeper in 6 states starting January 2026

The Trump administration will launch an AI program to approve or deny Medicare care, running January 2026 through 2031 across six states. Unlike current systems with human final decisions, this elevates AI to gatekeeper status for millions of seniors.

The timing is notable: As private insurers face backlash for AI denials reaching 19%, the federal government tests similar automation for 65+ million Medicare beneficiaries. Experts warn the cost-saving focus could compromise care quality and create a two-tier system.

Read the announcement: Trump AI Medicare pilot →

Johns Hopkins cuts $1.4M in costs with AI—by reducing call volume just 3%

Johns Hopkins Medicine receives 3 million scheduling calls annually. A 3% reduction via AI-powered call deflection, online scheduling, and callback features saved $1.4 million.

The lesson: Small percentage improvements at scale deliver massive financial impact. Johns Hopkins' hybrid approach combining AI with complementary tools rather than full automation suggests human-AI models work better than automation alone.

Read the case study: Johns Hopkins modernization →

Healthcare's AI arms race is here:

• Insurers weaponize: 500 AI tasks, 50K employees deployed, 19% denial rate, zero transparency

• Patients counter-attack: 1,000+ denials overturned, 6,000+ appeals generated, 25% of under-30s using AI for health advice

• States regulate: 4 ban AI-only decisions, dozen+ pass AI healthcare laws in 2025

• Government tests scale: AI Medicare gatekeepers for 65M+ beneficiaries starting Jan 2026

• Incrementalism wins: 3% improvement = $1.4M savings—scale beats moonshots

The future: Robots arguing with robots over whether humans get care. Progress or dystopia? Nobody's sure yet.

Talk soon — Avaamo Team