Skip to content
Search

Latest Stories

Top Stories

Why ChatGPT’s ‘memory’ will be a health care gamechanger

Woman looking at her smartwatch while exercising

Generative AI will have a pround impact on U.S. medicine writes Pearl. For example, home wearable devices will be able update to patients about their health status and suggest medication adjustments or lifestyle changes.

Guido Mieth/Getty Images

Pearl is a clinical professor of plastic surgery at the Stanford University School of Medicine and is on the faculty of the Stanford Graduate School of Business. He is a former CEO of The Permanente Medical Group.

OpenAI generated massive media interest with the announcement that its signature product, ChatGPT, is gaining memory. The new feature enables the generative artificial intelligence system to “carry what it learns between chats, allowing it to provide more relevant responses,” according to the company.

As Congress holds hearings and regulators rumble with apprehension, the media coverage so far has generally overlooked the biggest part of this announcement, which has direct ties to American health care:

The development of memory-powered AI is a pivotal step toward transforming U.S. medicine.


Although there are many technological and regulatory hurdles to clear — and fears around privacy and security to mitigate — this development has the potential to make health care more personalized, patient-centric and affordable. These improvements — alongside the potential pitfalls of AI-empowered health care — are the subject of my upcoming book, “ ChatGPT, MD: How AI-Empowered Patients & Doctors Can Take Back Control of American Medicine.”

Here are three ways generative AI’s improved memory will transform patient care:

More accurate diagnoses

For over a decade, clinicians have wanted to precisely tailor care to each patient’s unique health profile, including their genetic makeup and personal health preferences. But too much has stood in the way.

One major challenge is the sheer volume of knowledge required to customize medical care. The human genome consists of approximately 3 billion base pairs of DNA, which if typed out as letters would fill about 200 New York City phone books. What’s more, medical knowledge doubles every 73 days, making it almost impossible for any human to keep up with all the innovative medical findings and updated guidelines for helping patients.

A third hurdle is technological. With the average patient consulting 19 different doctors throughout their lifetime, an individual’s electronic medical records are often dispersed across numerous medical offices and health systems. The lack of interoperability among EMR systems compounds this issue, preventing clinicians —and, by extension, generative AI — from accessing a patient’s complete medical history.

Currently, ChatGPT’s “ context window ” (how many words it can recall before losing its memory), falls well short of the nearly 17,000 words found in the average patient’s medical record.

However, generative AI systems are predicted to become 30 times more powerful within the next five years, dramatically expanding their data retention capabilities and enhancing their reliability. This, combined with OpenAI’s specialized plug-ins (known as GPTs), offers promising opportunities. Initially, generative AI might access a limited set of patient data through platforms like My Chart, which can be used on personal computers or smartphones. Eventually, however, generative AI will enable patients to consolidate their digital medical records from various health care providers.

This will create a comprehensive, personalized health record, serving as a reliable resource for both patients and their health care teams.

With this information stored in an AI’s memory, patients will be able to input their symptoms and receive specific diagnostic and treatment suggestions.

For people who are uncertain about the significance or urgency of new symptoms, the AI will provide reliable advice. And for patients with rare or complex conditions, it will offer invaluable second opinions. Advanced diagnostic ability, alongside comprehensive health care information, will be instrumental in reducing the 400,000 annual deaths attributed to misdiagnoses.

Fewer complications from chronic sisease

Chronic diseases like diabetes, hypertension, obesity and asthma affect six in 10 U.S. adults. Complications from these diseases account for 1.7 million deaths each year.

Unlike acute illnesses that appear suddenly and usually are resolved quickly, chronic conditions persist over time, impacting tens of millions of Americans every single day.

Doctors care for these conditions in an episodic fashion, which is far from optimal. Patients with chronic diseases typically see their physician every three to four months, providing doctors with only a snapshot of their health status. As a result, chronic diseases aren’t controlled as well as they should be, which leads to life-threatening, and preventable, complications.

At a national level, hypertension is adequately controlled just 60 percent of the time, and effective blood sugar management in type 2 diabetes is achieved less than half the time. Data from the Centers for Disease Control and Prevention indicate that proper disease prevention and management approaches would reduce the risk of kidney failure, heart attacks and strokes by 30 percent to 50 percent.

Applying these percentages to the U.S. death toll from chronic disease complications, these CDC estimates indicate that more than half a million lives could be saved annually.

Generative AI, once connected to home wearable devices, can update patients about their health status and suggest medication adjustments or lifestyle changes. It can also remind them about necessary screenings and even facilitate testing appointments and transportation, thereby improving disease management, reducing complications and maximizing health outcomes.

Safer hospitals

Generative AI with memory will radically improve inpatient care, as well. Once it’s integrated with bedside monitors and able to remember a patient’s clinical status over time, the AI system will be able to immediately alert professionals when a problem arises, so they can intervene.

Additionally, video monitoring systems powered by AI could oversee the delivery of medical care, pinpointing any departures from established best practices. This real-time oversight would provide immediate alerts to caregivers, preventing medication mishaps and reducing the risk of infection.

These two uses of AI technology would help reduce the staggering 250,000 deaths each year attributed to preventable medical errors.

While ChatGPT and similar technologies hold immense potential, today’s generative AI tools still require clinician supervision. But looking ahead, the exponential growth of generative AI’s capabilities (doubling every year) points to a transformative future for the practice of medicine.

Now is the time for both clinicians and patients to become comfortable using generative AI. And it is an opportunity for regulators and elected officials to advance, not stifle its potential. With memory and GPTs, the doctor’s AI toolkit is quickly filling up.

Read More

Entertainment Can Improve How Democrats and Republicans See Each Other

Since the development of American mass media culture in the mid-20th century, numerous examples of entertainment media have tried to improve attitudes towards those who have traditionally held little power.

Getty Images, skynesher

Entertainment Can Improve How Democrats and Republicans See Each Other

Entertainment has been used for decades to improve attitudes toward other groups, both in the U.S. and abroad. One can think of movies like Guess Who's Coming to Dinner, helping change attitudes toward Black Americans, or TV shows like Rosanne, helping humanize the White working class. Efforts internationally show that media can sometimes improve attitudes toward two groups concurrently.

Substantial research shows that Americans now hold overly negative views of those across the political spectrum. Let's now learn from decades of experience using entertainment to improve attitudes of those in other groups—but also from counter-examples that have reinforced stereotypes and whose techniques should generally be avoided—in order to improve attitudes toward fellow Americans across politics. This entertainment can allow Americans across the political spectrum to have more accurate views of each other while realizing that successful cross-ideological friendships and collaborations are possible.

Keep ReadingShow less
Congress Must Not Undermine State Efforts To Regulate AI Harms to Children
Congress Must Not Undermine State Efforts To Regulate AI Harms to Children
Getty Images, Dmytro Betsenko

Congress Must Not Undermine State Efforts To Regulate AI Harms to Children

A cornerstone of conservative philosophy is that policy decisions should generally be left to the states. Apparently, this does not apply when the topic is artificial intelligence (AI).

In the name of promoting innovation, and at the urging of the tech industry, Congress quietly included in a 1,000-page bill a single sentence that has the power to undermine efforts to protect against the dangers of unfettered AI development. The sentence imposes a ten-year ban on state regulation of AI, including prohibiting the enforcement of laws already on the books. This brazen approach crossed the line even for conservative U.S. Representative Marjorie Taylor Greene, who remarked, “We have no idea what AI will be capable of in the next 10 years, and giving it free rein and tying states' hands is potentially dangerous.” She’s right. And it is especially dangerous for children.

Keep ReadingShow less
Microphones, podcast set up, podcast studio.

Many people inside and outside of the podcasting world are working to use the medium as a way to promote democracy and civic engagement.

Getty Images, Sergey Mironov

Ben Rhodes on How Podcasts Can Strengthen Democracy

After the 2024 election was deemed the “podcast election,” many people inside and outside of the podcasting world were left wondering how to capitalize on the medium as a way to promote democracy and civic engagement to audiences who are either burned out by or distrustful of traditional or mainstream news sources.

The Democracy Group podcast network has been working through this question since its founding in 2020—long before presidential candidates appeared on some of the most popular podcasts to appeal to specific demographics. Our members recently met in Washington, D.C., for our first convening to learn from each other and from high-profile podcasters like Jessica Tarlov, host of Raging Moderates, and Ben Rhodes, host of Pod Save the World.

Keep ReadingShow less
True Confessions of an AI Flip Flopper
Ai technology, Artificial Intelligence. man using technology smart robot AI, artificial intelligence by enter command prompt for generates something, Futuristic technology transformation.
Getty Images - stock photo

True Confessions of an AI Flip Flopper

A few years ago, I would have agreed with the argument that the most important AI regulatory issue is mitigating the low probability of catastrophic risks. Today, I’d think nearly the opposite. My primary concern is that we will fail to realize the already feasible and significant benefits of AI. What changed and why do I think my own evolution matters?

Discussion of my personal path from a more “safety” oriented perspective to one that some would label as an “accelerationist” view isn’t important because I, Kevin Frazier, have altered my views. The point of walking through my pivot is instead valuable because it may help those unsure of how to think about these critical issues navigate a complex and, increasingly, heated debate. By sharing my own change in thought, I hope others will feel welcomed to do two things: first, reject unproductive, static labels that are misaligned with a dynamic technology; and, second, adjust their own views in light of the wide variety of shifting variables at play when it comes to AI regulation. More generally, I believe that calling myself out for a so-called “flip-flop” may give others more leeway to do so without feeling like they’ve committed some wrong.

Keep ReadingShow less