Can we deliver high-quality healthcare without AI and automation?

Karandeep Singh, MD, MMSc

Joan & Irwin Jacobs Chancellor’s Endowed Chair and Associate Professor of Medicine

Executive Director, Jacobs Center for Health Innovation

Chief Health AI Officer, UC San Diego Health

Can we deliver high-quality

healthcare without AI and

automation?

Disclosures

• Participate on Google’s Consumer Health Advisory Panel as a paid consultant

−Involved in FitBit, not the Articulate Medical Intelligence Explorer (AMIE)

• Participate on the BMJ Digital Health Advisory Board as a paid consultant

2 │

What does American healthcare look like today?

“A recent faculty recruit in my department, an expert in artificial intelligence (AI), asked me whether I could

recommend a primary care doctor. I assured my colleague that I would promptly provide a set of excellent

options to choose from. When I reached out to several doctors who were my contemporaries in medical

training, however, I learned that they had all retired. That was not completely surprising, so I tried another

tack: I contacted doctors who were of more recent vintage and whom I’d encountered in a supervisory

capacity. One of them cheerfully informed me that he was no longer practicing primary care at a teaching

hospital affiliated with Harvard Medical School. When I asked him whether he could recommend any high-

quality practices, his tone changed, and he said that he couldn’t think of any such practices that were

accepting patients. Stunned, I widened my search further. I discovered that primary care clinics at some local

hospitals, such as Massachusetts General Hospital and Brigham and Women’s Hospital, were turning away

new patients. I did not end up helping my new colleague and instead suggested contacting their health

insurance company for recommendations.”

3 │

What does high-quality healthcare look like?

4 │

AI and automation offer an approach that could help

https://www.ama-assn.org/practice-management/cpt/cpt-appendix-s-taxonomy-artificial-intelligence-medical-services-procedures

5 │

Assistive AI

Output provides clinically relevant data

without deriving a parameter (index,

score, classification), and offers no

interpretation or conclusion.

Augmentative AI

Output is a quantitative or categorical

parameter qualitatively different from the

input, more than adding, averaging, or

reporting descriptive statistics.

Autonomous AI

Software independently derives

parameters and generates clinically

meaningful interpretations without

concurrent physician or QHP involvement.

AUTONOMOUS AI: LEVELS OF OVERSIGHT

LEVEL I

Recommends a definitive diagnosis or

a specific intervention. A physician or

QHP must judge whether to implement

or reject it.

LEVEL II

Recommends medical management,

with a reasonable opportunity to negate

the impending action before

implementation (e.g., an alert).

LEVEL III

Initiates management actions

automatically and continues under

physician oversight unless a clinician

chooses to intervene.

Can AI improve

population health?

Intervention started

Deaths (relative vs. expected)

Augmentative AI can save lives in hospitalized patients

COMPOSER

(Boussina et. al. 2024)

TREWS

(Adams et. al. 2022)

Can AI improve the

experience of care?

Do patients trust health systems to use AI?

Nong P, Platt J. Patients’ Trust in Health Systems to Use Artificial Intelligence. JAMA Netw Open. 2025;8(2):e2460628. doi:10.1001/jamanetworkopen.2024.60628

9 │

“This analysis found low trust in health care systems to

use AI responsibly and protect patients from AI-related

harms. General trust in the health care system, but not

health literacy or AI knowledge, was associated with

these perceptions.”

Patients regularly use AI to ask health questions…

10 │

…despite their lack of trust in the information

11 │

But they trust when health systems actually use AI

12 │

Colonoscopy & Endoscopy AI Preparatory Calls

Launched on December 11, 2025

• Over 10,000 calls made

• “This was the smoothest AI involvement I have

ever had with a phone call, and I usually prefer

talking to a regular, real, normal human being.”

What about patient trust in AI in the real world?

13 │

Key Metrics to Highlight Baseline Target To-Date

No-Show & Same-Day Cancel. 10-15% 5%* 6%**

Attempt Outreach 36.2% 100% 100%*

Connected Rate 24.6% 67%

*Benchmarked against top-performing ASCs with extensive outreach operations.

**6% amongst all attempted patients; 4% for patients that complete their call.

***La Jolla and Hillcrest (~3,000 calls for 1,600 appointments). East Campus next.

Unpublished.

Can AI improve

workforce

wellbeing?

Improved wellbeing finding is immediate and sustained

15 │

UWisconsin

Afshar et al. NEJM AI 2025

Can AI reduce

costs?

Use of AI CPT codes has grown 10-fold over the past 5 years

Unpublished. In collaboration with Drs. Eileen Kim and Mattheus Ramsis.

17 │

To reduce costs, we need to learn from other industries

18 │

How can we achieve cost savings with AI?

19 │

Can AI improve

health equity?

Patient use of AI has upended the digital divide

21 │

What’s next in AI?

Autonomous AI is entering the space of clinical care

23 │

Autonomous AI is being compared against physicians…

Note: This is a preprint.

24 │

…but that only matters if there are physicians available…

25 │

…and if those physicians can see you in a reasonable timeframe

26 │

Mental Health

Can we deliver high-quality healthcare w ithout AI & automation?

27 │

What role should autonomous AI play in our health system?

Status quo

Change nothing

Expand the

workforce

Delegate to other

health professionals

AI supports

physicians

Example: AI scribes

Same panel size

AI extends

physicians

Example: ARPA-H

ADVOCATE

Larger panel size

Physician out

of the loop

Example: Utah AI

medication refills

Increasing reliance on AI →

28 │

Which one would you pick?

(with physicians as an example)

Jeffrey Pan

Co-Director

JCHI

Keisuke Nakagawa, MD

Director of Strategic Impact

JCHI

Joshua Glandorf

Chief Information Officer

UC San Diego Health

Nicole May

Co-Director

JCHI

Karandeep Singh, MD, MMSc

Executive Director, JCHI

Chief Health AI Officer, UCSD Health

Jarrod McDonald

Chief of Staff & Sr. Director

CMO Administration

Alex Khalessi, MD, MMSc

Chief Innovation Officer

UC San Diego Health

Thank you