In our last blog, we explored the idea that AI is accelerating so rapidly, we may need a new way to measure its progress—akin to Moore’s Law, but faster. I shared this on LinkedIn as:
Are We Ready for “AI Years™”?
I was playing with my aging dog when the thought struck me: just as we use “dog years” to capture the accelerated aging of our pets, maybe we need “AI Years” to describe the breathtaking pace of artificial intelligence.Moore’s Law—Gordon Moore’s observation that transistor counts double approximately every two years—guided the cadence of innovation in computing for decades. But AI is moving at a different, more exponential pace.
That’s where “AI Years™” comes in.
One AI Year is a conceptual time unit—often 3 to 6 months—during which AI capabilities double or leap ahead in performance. We’re not just seeing bigger models, but smarter reasoning, better language understanding, improved decision-making, and real-world applications that didn’t even exist months ago.
1 AI Year ≈ 6 calendar months, depending on the domain and the innovation cycle.
This framework helps explain why institutions—product teams, regulators, educators, policymakers—struggle to keep pace. They move in traditional years. AI evolves in AI Years.
The question is no longer if AI is moving fast. It’s whether we are moving fast enough to keep up.
The Healthcare AI Adoption Index: Progress, But Not Without Friction
In Part I of this series, we referenced Bessemer Venture Partners®’ Healthcare AI Adoption Index, developed in partnership with AWS® and Bain & Company®. The report draws on insights from over 400 healthcare buyers, identifying what drives AI experimentation and the challenges of moving from proof-of-concept to production.
One key theme? AI adoption in healthcare is a “fast-slow” process (our words, not theirs). Projects often start with high energy but stall at implementation. Despite this, some successful use cases are emerging—especially where startups partner directly with hospitals or payors. But these are the exceptions, not the norm.
Consider ownership structures: while some hospital systems own their own insurance plans (especially for employees), very few payors own hospitals. Kaiser Permanente stands out as a fully integrated model. UnitedHealth, meanwhile, operates across the full spectrum—insurance, technology, and care delivery—making them a $350 billion market cap company. But again, these are outliers.
Most of the time, payors, providers, and hospitals operate with fundamentally different incentives. When someone claims their innovation will “cut costs” or “improve quality,” it’s critical to remember: in healthcare, one party’s cost savings can mean lost revenue for another. Better care often costs more, and who bears that cost is always up for debate.
The Real Challenge: Injecting Innovation Into Complex Systems
Injecting innovation—especially AI—into existing healthcare systems is extraordinarily difficult. From personal experience, I’ve found that change is most successful when it originates from within. Top-down mandates rarely stick.
Some healthcare systems are starting to “suggest” using AI, partly to train it. But adoption faces major resistance. As this Bessemer slide shows, 55% of large healthcare organizations struggle with behavioral change around AI. No one wants to risk their job by relying on technology that’s still learning on the job.
My Conflict with Direct-to-Patient AI in Healthcare
I’m a strong advocate for AI that improves the consumer healthcare experience. But I remain cautious about AI’s current utility within the clinical system. We need AI tools that are faster, better, and more accurate than humans—not ones that merely add more work for already overstretched providers.
The Concern
AI is undoubtedly changing healthcare. Fields like radiology, pathology, and genomics are already benefiting. But challenges remain. Consider this recent research finding:
“GenAI-drafted replies were associated with significantly increased read time, no change in reply time, significantly increased reply length, and some perceived benefits. Rigorous empirical tests are necessary to further examine GenAI’s performance.”
Or this physician’s frustration with Epic’s ART system:
“It uses generic, one-size-fits-all system prompts that users cannot access. The result? Most drafts are useless.”
And in one joint Stanford-University of Colorado study, clinicians used only 12% of AI-generated responses in full, and partially used another 20%. The rest? Discarded.
AI today is largely driven by FOMO—fear of missing out—and the search for efficiency gains. But in most cases, AI cannot outperform providers in treating real patients, at least not yet.
We Are Still Very Early—Despite the Hype
Despite the dizzying pace, we’re in the very early stages of AI in healthcare. The competition is global, with the U.S., China, and Europe all jockeying for dominance. News in this space breaks not weekly, not daily—but hourly. Below is a final relevant slide from Bessemer’s report.
AI Is “Perfect Enough” for the Consumer—Today
As a company, and as someone with 30 years of healthcare experience, one thing is clear: AI is moving fast in AI Years™, but it’s still not ready to help most doctors.
If using AI today means doctors must spend even more time correcting, editing, or prompting, then it won’t work—not today. That’s why as many as 99% of AI healthcare projects may fail. The bar is higher in healthcare than anywhere else:
-
1,000,000+ doctors
-
1,000+ insurance companies
-
700,000+ ERISA employer plans
-
5,000+ acute care hospitals
-
850 million+ annual office visits
-
6+ billion annual prescriptions
-
10+ billion medical claims
-
340 million+ U.S. consumers
The opportunity—and the complexity—of AI in healthcare is staggering. We already have:
-
Over 1.2 billion medical records
-
Over 500 EHR systems
-
More than 3,000 exabytes of data
Yet the consumer remains the most motivated, the most engaged, and the one with the most to gain. A thoughtfully built, AI-powered health assistant—curated, secure, and easy to use—could offer real benefits today.
Until AI is good enough to operate without costly supervision from clinicians (whose time runs $10–$30 per minute), its most practical role is to empower the consumer—not the provider.
About HealthScoreAI ™
Healthcare is at a tipping point, and HealthScoreAI (HSAI) is positioning to revolutionize the industry by giving Consumers control over their health data and unlocking its immense value. U.S. healthcare annual spending has exceeded $5 trillion with little improvement in outcomes. Despite advances, technology has failed to reduce costs or improve care. Meanwhile, 3,000 exabytes of Consumer health data remain trapped in fragmented USA system of 500 EHRs, leaving Consumers and doctors without a complete picture of care.
HealthScoreAI seeks to provide a unique solution, acting as a data surrogate for Consumers and offering an unbiased holistic view of their health. With over 850 million medical claims denied annually in the U.S., HSAI intends on giving Consumers practical tools to respond to denial of care by insurers. We aim to bridge the gaps in healthcare access and outcomes. By monetizing de-identified data, HealthScoreAI seeks to share revenue with Consumers, potentially creating a new $100 billion market value opportunity. With near-universal EHR adoption in the USA, and advances in technology, now is the perfect time to capitalize on the data available, practical use of AI and the empowering of Consumers, in particular the 13,000 tech savvy baby boomers turning 65 every single day and entering the Medicare system for the first time. Our team, with deep healthcare and tech expertise, holds U.S. patents and a proven track record of scaling companies and leading them to IPO.
Noel J. Guillama-Alvarez
https://www.linkedin.com/in/nguillama/
+1-561-904-9477, Ext 355