1 of 28

Current & Future State of AI

Shivani Nazareth, MS, CGC

VP, Digital Health Strategy at Myriad Genetics

2 of 28

No conflicts or disclosures to declare

Nazareth | Dec 2024 | NYSGTF | Confidential

3 of 28

AI Fatigue: Decoding the Noise

Nazareth | Dec 2024 | NYSGTF | Confidential

4 of 28

The Evolution of AI

ARTIFICIAL INTELLIGENCE

Machines that mimic human intelligence

MACHINE LEARNING

Algorithms that learn without explicit training

DEEP LEARNING

Subset of machine learning that uses large data sets

GENERATIVE AI

AI that creates new content from data

Nazareth | Dec 2024 | NYSGTF | Confidential

5 of 28

2023 – a pivotal year for AI technology

Nazareth | Dec 2024 | NYSGTF | Confidential

Generative AI

6 of 28

One year later…

AI and chatGPT studies are published across various specialties of medicine; 10K/month

Nazareth | Dec 2024 | NYSGTF | Confidential

7 of 28

Nobel Prize Winners in Physics 2024

They used physics to find patterns in information and lay the foundation for machine learning.

Nazareth | Dec 2024 | NYSGTF | Confidential

8 of 28

Google vs LLM

Nazareth | Dec 2024 | NYSGTF | Confidential

PROMPT: Can you suggest a 4-day itinerary in Paris on a $2,000 budget?

9 of 28

AI will reduce time to diagnosis

Nazareth | Dec 2024 | NYSGTF | Confidential

Making sense of structured and unstructured data:

AI-enhanced EHRs have the potential to help clinicians scan medical records (structured and unstructured) holistically and display diagnostic possibilities much faster than any individual provider.

10 of 28

JAMA viewpoint: Sept 2023

Nazareth | Dec 2024 | NYSGTF | Confidential

“Yesterday I saw a child with [a] variety of features characteristic of a genetic syndrome. I asked to chatGPT for a diagnosis and it came out within seconds with Rubinstein-Taybi syndrome. Astounding and overwhelming. Further, it suggested all the tests I would need, putting cAMP-response element binding protein (CREBBP) genome assay as the most desired. The child at a very young age has been seen by a geneticist at a tertiary center and the tests done did not include CREBBP genome sequence. The technology is a godsend for rural physicians like me because the nearest tertiary center is about 90 miles away and rural America is like the developing world within America.” – Dr. Dave, pediatrician in rural IL

11 of 28

AI versus MD

“In this cross-sectional study of human-written and AI-generated responses to 200 eye care questions from an online advice forum, a chatbot appeared capable of responding to long user-written eye health posts and largely generated appropriate responses that did not differ significantly from ophthalmologist-written responses in terms of incorrect information, likelihood of harm, extent of harm, or deviation from ophthalmologist community standards.”

Nazareth | Dec 2024 | NYSGTF | Confidential

12 of 28

“AI will not take your job, but people who know how to use AI will.”

Dr. Matt Wood

VP for AI Products at Amazon (AWS)

AWS Summit New York 2024

Nazareth | Dec 2024 | NYSGTF | Confidential

So where does that leave us? I attended the Amazon Summit in NYC earlier this year; I spoke on a panel of people from various industries using AI. In my case, I talked about how my employer is using AI to help compile the information that insurance companies need to cover the cost of testing. This is experimental, but if we can figure it out, we can process claims faster. One of the other members of the panel was the head of technology at Ralph Lauren– he talked about how they are using AI to make it easier to exchange and return customer purchases. What I took away is that companies across all sectors, not just healthcare, are trying to figure out how to integrate AI into their workflow.

Matt Wood, who led AI at Amazon said during his keynote talk: AI will not take your job, but people who know how to use AI will. I am not worried that AI will take a genetic counselor’s job, but I think it’s fair to say that a genetic counselor who is willing to use AI is more likely to get the job in an interview. Every hospital system is looking for ways to make healthcare more efficient.

13 of 28

The internet was a disruptive technology

2017: Next View

Nazareth | Dec 2024 | NYSGTF | Confidential

14 of 28

October 2024:

FDA announced the creation of a new Digital Health Advisory Committee to help it explore scientific and technical issues related to digital health technologies including AI.

FDA has approved nearly 950 AI-enabled devices as of October 2024.

Modern Healthcare | fda.gov

AI-enabled medical devices

Image Source: https://medicalfuturist.com/the-current-state-of-fda-approved-ai-based-medical-devices/

Nazareth | Dec 2024 | NYSGTF | Confidential

15 of 28

“There’s no question we have issues with access to care, and whether or not it is a good idea to deploy ChatGPT to cover the holes or fill the gaps in access, it’s going to happen and it’s happening already…”

“People have already discovered its utility. So, we need to understand the potential advantages and the pitfalls.”

Dr. Nieraj Jain, assistant professor at Emory

AI and EMR/EHR

Nazareth | Dec 2024 | NYSGTF | Confidential

16 of 28

“This isn’t science fiction.”�

"Generative AI lets your teams get through everyday repetitive tasks and spend more time on things that bring them meaning. And saving a little time here and there — that really adds up," Epic said in an Oct. 13 video. "This isn't science fiction…”

Nazareth | Dec 2024 | NYSGTF | Confidential

17 of 28

Epic approves Abridge as a vendor

Abridge’s software uses generative AI to convert a patient-clinician conversation into a structured clinical note draft. Clinicians then review each draft note before inputting it into the electronic health record.

https://www.modernhealthcare.com/digital-health/abridge-ai-mayo-clinic-cvs-health-ventures-epic

Nazareth | Dec 2024 | NYSGTF | Confidential

18 of 28

Amazon→ Healthcare

Amazon Web Services (AWS) has rolled out “HealthScribe”, a pioneering clinical documentation service rooted in generative AI. This platform facilitates the creation of clinical applications by utilizing speech recognition and AI to transcribe patient encounters, discern critical details, and craft concise summaries apt for electronic health record (EHR) integration.

https://medicalfuturist.com/amazon-in-healthcare-disruption-data-nightmares-and-digital-health-dreams/

Nazareth | Dec 2024 | NYSGTF | Confidential

19 of 28

Ambient transcription

https://www.mckinsey.com/industries/healthcare/our-insights/tackling-healthcares-biggest-burdens-with-generative-ai

Clinician downloads a HIPAA-compliant AI app

App can listen to clinician-patient conversation and act as a scribe

App generates real-time notes using conversational language

After the visit, the clinician can review note and fill in any gaps

Notes can be submitted to EHR

Patient follow-up care and/or next steps/discharge summary can be printed in any language

Nazareth | Dec 2024 | NYSGTF | Confidential

20 of 28

Health insurance prior-auth

Health insurance prior authorization and claims processing takes an average of 10 days. AI-enabled algorithms can be used to generate real-time benefits verification and calculate out-of-pocket costs, using contracted rates + the patient’s specific benefits. This would be in lieu of calling the number on the back of the insurance card.

Source: The AI Revolution in Medicine: GPT-4 and Beyond ; by Peter Lee, Carey Goldberg, Isaac Kohane ; Publisher(s): Pearson ; ISBN · 9780138200145. Image from Decoderize

Nazareth | Dec 2024 | NYSGTF | Confidential

21 of 28

AI capabilities that would improve healthcare

AI for identifying rare disease variants

Remote patient monitoring using wearable devices

Auto-generating a letter of medical necessity and faxing it to the insurance company

Data-mining medical records and flagging patients who match eligibility for clinical trials

Augmented reality views for optimizing surgical procedures

Nazareth | Dec 2024 | NYSGTF | Confidential

22 of 28

What’s Next?

Artificial Narrow Intelligence

Artificial General Intelligence

Artificial Super

Intelligence

“First announced in a 2017 paper from Google researchers, transformers are a kind of AI algorithm that lets computers understand the underlying structure of any heap of data—be it words, driving data, or the amino acids in a protein—so that it can generate its own similar output.”

Nazareth | Dec 2024 | NYSGTF | Confidential

23 of 28

Contextual understanding

“That level of contextual understanding enables transformer-based AI systems to not only recognize patterns, but predict what could plausibly come next—and thus generate their own new information. And that ability can extend to data other than words.” - WSJ Nov 2024

WAYMO

Nazareth | Dec 2024 | NYSGTF | Confidential

24 of 28

A fun listen

We’ve witnessed a revolution in A.I. since the public rollout of ChatGPT. Our Senior Editor David Kestenbaum thinks that even though there’s been a ton of coverage, there’s one thing people haven’t talked much about: have these machines gotten to the point that they’re starting to have something like human intelligence? Where they actually understand language and concepts, and can reason? He talks with scientists at Microsoft who’ve been trying to figure that out. (30 minutes)

Nazareth | Dec 2024 | NYSGTF | Confidential

25 of 28

"[Evolutionary algorithms in AI] are inspired by Charles Darwin’s work on natural selection. Darwin discovered that the strongest members of a species survive over time, and their genetic code goes on to dominate the population. Over time the species become better suited to its environment.

So it is with artificial intelligence. Initially, a system starts with a very large semi-random set of possibilities…and runs simulations…The system will strip out the weak and keep the strong and then create a new combination. …Evolutionary algorithms with the power to mutate will help advance AI on its own…”

AI will self-evolve

Nazareth | Dec 2024 | NYSGTF | Confidential

26 of 28

AI regulation�

"Another thing is clear: the US presidential election in 2024 will color much of the discussion on AI regulation. As we see in generative AI’s impact on social media platforms and misinformation, we can expect the debate around how we prevent harms from this technology to be shaped by what happens during election season.”

Nazareth | Dec 2024 | NYSGTF | Confidential

27 of 28

Coalition for Health AI

Nazareth | Dec 2024 | NYSGTF | Confidential

28 of 28

“If one were to tell an unborn child that outside the womb there is a glorious world with green fields and lush gardens, high mountains and vast seas, with a sky lit by the sun and the moon, the unborn would not believe such absurdity.”

Poetry by Rumi, 1207-1273

Nazareth | Dec 2024 | NYSGTF | Confidential