Using AI to Scale the World’s Best Healthcare to Every Human Being

Neal Khosla
May 30, 2018 · 5 min read

Imagine if you could text with an AI doctor as if you were asking questions to your human physician. You would be able to send information about your symptoms, pictures of your skin rash; or perhaps images of your child’s bleeding cut to see if she needs stitches or a visit to the emergency room; or receive blood test orders or results directly. The AI would know your medical history, and it would have precise and up-to-date information about the prevalence of diseases in your area, as well as the latest research on treatments and their efficacy. And unlike your typical physician, it would be available 24x7. The AI would be smart enough to know when it doesn’t know enough, and would triage up to the right level of human healthcare professional. Now imagine if this personal AI doctor was available as a primary care physician to all 7 billion people on this planet.

This is our mission at Curai — to We believe that great healthcare is a human right and this is the future we’ve set out to build.

The Team

Curai was founded by an experienced team: Neal Khosla, who combines passion about this problem with an ML background from Stanford and Google; Xavier Amatriain, previously VP of Engineering at Quora and before that leading the team that built the famous Netflix recommendation algorithms; and Neil Hunt, formerly Chief Product Officer at Netflix. We have already hired some amazing people, and we are looking to dramatically expand resources — with the very best talent in AI, product management, and operations — to scale solutions to this problem.

Why Primary Care?

Because every person interacts with primary care as their entry point into the healthcare system, it is the only place where integrative care across multiple diseases like cardiac care, diabetic care, mental health care can happen. Although primary care is not expensive itself, it directs a lot of healthcare spend and hence can be a significant driver of cost to insurance companies and others who bear healthcare spending risk.

In the US, access to quality primary care is a real problem, with average wait time of 24 days just to see a doctor and 13% of Americans not able to see a physician because of cost. This is only expected to get worse with a growing shortage of trained doctors (expected to grow to 100k+ in the US by 2030). Globally, where many people don’t have any access to a doctor at all, the problem is even worse. The exponential increase in medical knowledge (more than 200 journal articles published daily) makes it hard for doctors to stay current. All of these facts suggest to us that a dynamic AI/ML system will solve this fundamental access problem and grant this universal right to the world’s population.

Patient Centered, Physician in the Loop

We believe that the patient is the most important part of any health care interaction. Patients too often get thrown by the wayside in the world of documentation, complex hospital systems, and confusing terminology. We are using our experience in creating consumer products to create tools that patients can use to truly understand their health concerns and issues. In the modern healthcare system, physicians have lost much of their ability to focus directly on the patient. We want to bring this patient-centric universe back.

A key part of our approach is centered around facilitating the interaction between a patient and a physician. We want to ensure every physician has the information they need about their patients. As Robert Wachter of USCF notes in his book, The Digital Doctor:

“If we go into the visit with both parties knowing what’s going on, then we can spend the visit talking about what we do about it.”

If this interaction happens in the right way, patients can regain ownership of their medical decisions and, of course, their data.

Data and Machine Learning will Enable the Science of Medicine

A tenet that permeates our culture at Curai is a belief in the power of data to transform medical decision making into an evidence-backed process. While there is a lot of debate about what “evidence-based” medicine actually looks like, it is clear that as an industry, medicine has a long way to go. Data-driven decision making and AI have already transformed industries as broad as finance, media, and sports, so we (and many others) strongly believe it should change the way we make medical decisions. There are many examples of medical decisions without sufficient data to support them (see for example, this study, that found that nearly 90% of ACA based clinical guidelines in cardiology were not based on Grade A clinical evidence). Atul Gawande eloquently gets at the heart of this in his book, Being Mortal:

“We look for medicine to be an orderly field of knowledge and procedure… There is science in what we do, yes, but also habit, intuition, and sometimes plain old guessing. The gap between what we know and what we aim for persists.”

We want to help close that gap. We know it’s something that many great researchers have worked on. That’s why we have begun a collaboration with Nigam Shah’s laboratory at Stanford and are being advised by Professor David Sontag outside of his capacity at MIT. We’ve also licensed some of the best prior work in the field including the VDDx system built by Randy Miller and others at Vanderbilt University and the DXplain system built by the talented folks at the Laboratory of Computer Science at Massachusetts General Hospital.

We are bringing our own flavor of decision making backed by machine learned models that are validated and enhanced by the knowledge of physicians. Check out our work on combining learned models with expert systems, and our CTO Xavier’s talk about the challenges in the space and interesting approaches to solve them.

Our Journey

At Curai, we believe that every human being deserves access to world-class healthcare. We know that these are incredibly complex and nuanced problems that won’t be solved without great minds, irregular patience and persistence, and cross disciplinary collaboration. That’s why we’ve put together a world class team of physicians and machine learning experts from places like Stanford Medicine, Facebook AI Research, and Netflix.

That’s also why we joined forces with some of the best VC’s as we are pleased to announce our Series A, backed by General Catalyst, Khosla Ventures, and a variety of angels. Throughout our process of talking to folks, we were continually struck by the thoughtfulness, mission driven approach, and desire to take a long view of these firms and their partners as well as their expertise, resources, and connections in both the AI and healthcare worlds.

Early on, we made it our goal to get the brightest minds of today to stop working on optimizing selling ads and to come join us in our mission. If you’re the type of person who believes deeply in the power of data to change the way decisions are made, and who wants to be a part of bringing data driven healthcare to the world, come join us.


Applying AI/ML to improve access, efficiency, and outcomes for primary health care

Neal Khosla

Written by

Cofounder, @CuraiHQ. Cofounder, @TotemicLabs. Past associations w/ @YCombinator, @Stanford, @Google, @DormRoomFund



Applying AI/ML to improve access, efficiency, and outcomes for primary health care