Healthcare in the AI Era

Healthcare in the AI Era

This week, I want to look into the different ways Artificial Intelligence (AI) can be used in healthcare and how it could bring significant benefits for both patients and healthcare providers.

First, I'd like to review a short history of AI to explain why this moment is so important and highlight the key milestones that have shaped the industry.

Evolution of Artificial Intelligence

The Turing Test has been explained theatrically, but this framework can help us better understand how AI has evolved and where we are. Here is a summary of some of the key milestones where AI has achieved the same level of accuracy as humans or even surpassed them in some tasks.

  • 1950: Alan Turing proposed the Turing test, a criterion for judging whether a machine can exhibit intelligent behavior equivalent to or indistinguishable from a human. [1]
  • 1956: The term "artificial intelligence" was coined by John McCarthy at the Dartmouth Conference, where the foundations of AI research were laid. [2]
  • 1965: Joseph Weizenbaum created ELIZA, a natural language processing program that could simulate a psychotherapist and engage in a conversation with a human. [2]
  • 1997: IBM's Deep Blue defeated world chess champion Garry Kasparov in a six-game match, demonstrating the ability of AI to perform complex reasoning and strategy. [2]
  • 2011: IBM's Watson won the quiz show Jeopardy! against two former champions, showing the ability of AI to understand natural language, process large amounts of data, and generate accurate answers. [2]
  • 2014: Google's DeepMind developed AlphaGo, a program that beat the European Go champion Fan Hui, and later in 2016, the world Go champion Lee Sedol, demonstrating the ability of AI to learn from data and self-play and master a game that is more complex than chess.
  • 2015: Microsoft and Google achieved human parity in image recognition, meaning that their AI systems could identify objects in images as well as humans. [3]
  • 2016: Google's DeepMind developed WaveNet, a neural network that could generate realistic human speech, surpassing the quality of existing text-to-speech systems. [4]
  • 2017: Alibaba and Microsoft achieved human parity in machine reading comprehension, meaning that their AI systems could answer questions based on a given text as well as humans. [5]
  • 2020: OpenAI developed GPT-3, a deep learning model that could generate coherent and diverse texts on various topics and tasks, such as writing essays, creating chatbots, and composing code.

To start connecting the dots with the medical field, here are some milestones and measurements that show how AI is performing against doctors or other specialists:

  • 2017: Researchers from Stanford University developed a deep learning algorithm that could diagnose skin cancer from images of skin lesions with an accuracy comparable to that of dermatologists. [6]
  • 2018: Researchers from Google and its health-tech subsidiary Verily developed a deep learning algorithm that could detect diabetic retinopathy, a leading cause of blindness, from eye scans with an accuracy similar to that of ophthalmologists. [7]
  • 2023: Google Med-PaLM [8]

Earlier this year at C2 Montreal, I attended a presentation from Google, where the team showed some of their latest Gen AI releases; one of them was Med-PaLM, which is designed to provide high-quality answers to medical questions. [8]

Google has recently introduced its latest model, Med-PaLM 2, which achieves an accuracy of 86.5% on USMLE-style questions, a 19% leap over its own state-of-the-art results from Med-PaLM. Med-PaLM 2 also substantially improves the generation capabilities of large language models for producing long-form answers to consumer medical questions. [9]

Technology is evolving faster and faster, but in this field, I don't think that will be enough. There are still multiple challenges that have to be addressed on top of the regulatory and social processes. In a recent interview , Bill Gates shared his predictions on where AI is going. He emphasized that the next 2-5 years could bring significant increases in accuracy and reductions in development costs, and with those, generative AI could be viable for medical use cases like drug development and health advice.

Applying AI in the Healthcare Industry

The World Economic Forum's report on incorporating innovative AI solutions in health [10], produced with ZS, points out high-priority scenarios for teamwork between the public and private sectors in healthcare. It emphasizes the need for a shared vision and active steps to tap into the full potential of AI for improving global health. The report lists six key use cases that can make a big difference in health outcomes, access, and efficiency: 1) monitoring patients remotely; 2) virtual health helpers; 3) supporting clinical decisions; 4) advancing drug discovery and development; 5) preparing for and responding to pandemics; and 6) optimizing health systems.

Another article from the World Economic Forum about the impact of Generative AI in healthcare [11] explores how this type of AI can help tackle challenges in the healthcare sector, such as pandemics, chronic diseases, mental health issues, and a shortage of medical professionals. It discusses the potential of Generative AI in creating synthetic data, images, and sounds that can be used for diagnosis, treatment, and education. Additionally, it highlights how generative AI can help craft personalized treatment plans based on an individual's genetic, environmental, and lifestyle factors. The article also touches on enhancing patient engagement and satisfaction through virtual assistants, chatbots, and remote monitoring devices.

Exploring the economic impact of Artificial Intelligence (AI) in healthcare unveils a future full of promise. Various reports have tallied the potential savings and improvements, showcasing a scenario where AI not only makes healthcare more efficient but also boosts global economies and betters health outcomes. The numbers suggest a significant change is on the horizon, with substantial savings and improved life expectancy. Let's dive into the data to see the financial and health benefits AI could bring to the worldwide healthcare scene.

  • The report forecasts a bright economic outlook, with AI poised to trim down healthcare expenses worldwide by a hefty $400 billion annually by 2026 while also propelling the global GDP upwards by 14% come 2030 [12].
  • On the health frontier, the report is optimistic that AI will be a game-changer, envisaging a 10% dip in mortality rates, an extra four years tacked onto healthy life expectancy, and a staggering 2.5 million lives saved from the clutches of infectious diseases by 2030 [12].
  • Delving into the healthcare workforce, the report projects AI as a catalyst for heightened productivity, enabling a 50% surge in the patient count per health worker, a 30% reduction in time frittered away on administrative chores, and a 40% boost in the precision of diagnosis and treatment by 2030 [12].
  • Echoing a similar tune, the World Health Organization (WHO) report shares that AI could potentially pare down global healthcare costs by $400 billion annually by 2026, alongside a 14% ascension in global GDP by 2030 [13].
  • The IQVIA Consumer Health study2 sheds light on the consumer health sphere, where AI chatbots are seen as a lever to jack up consumer health engagement by 50%, pare down medication non-adherence by 30%, and enhance self-care outcomes by 20% [14].

Real Problems to Solve

A few weeks ago, someone close to our family was diagnosed with liver disease. By the time she decided to visit the hospital, her health had already worsened significantly. This experience, which unfolded in Mexico's public healthcare system, could unfortunately happen anywhere.

Usually, medical care involves brief consultations, sometimes as short as 15 minutes, despite long waiting times. And depending on the case, a high bill might come with it.

In the scenario I mentioned, likely, the symptoms started much earlier. But with many hurdles in the way, going to the hospital or doctor often becomes a last resort. And even then, the process is not straightforward, quick, or, in many instances, affordable.

According to a report by The Lancet Global Health Commission on High-Quality Health Systems [15], 5.7 million people die in low and middle-income countries every year due to poor-quality healthcare, compared to the 2.9 million who die from lack of access to care. This suggests that many people die because they can't get timely and effective treatment.

Survey data from Gallup shows that 25% of Americans say they or a family member have delayed medical treatment for a serious illness because of the costs. Another survey by Forbes reveals that 13% of U.S. adults, about 34 million people, know of at least one friend or family member who died in the last five years because they couldn't afford medical treatment [16]. These figures indicate that financial barriers often prevent people from getting the medical help they need.

Time to Build

AI isn't meant to replace doctors, but it can help them in many ways and improve patient experience.

With technology getting better and specialized models like Med-PaLM coming into the picture, there's much to look forward to.

I see a future where new services give patients early, easy, and affordable access to good information. This can help them make better decisions and find the right services when needed.

AI can also support doctors in far-off places and developing countries, helping them provide better care by boosting their skills.

There are real risks when it comes to mixing AI and medicine. We need good regulations in place to make sure healthcare products and services are safe and effective. Right now, the rules for AI and medicine are quite different; there is a lot of work in front of us.

The convergence of AI and medicine is really exciting to me. It's worth thinking about places without enough healthcare support and how AI could help. With AI, making the right decisions earlier, getting better diagnoses, and using data wisely could improve healthcare services everywhere.

This could be a big step towards making healthcare better and more accessible for everyone, no matter where they live.

Let's build.


Source: Conversation with Bing

[1] The brief history of artificial intelligence: The world has changed .... https://ourworldindata.org/brief-history-of-ai .

[2] History of artificial intelligence - Wikipedia. https://en.wikipedia.org/wiki/History_of_artificial_intelligence .

[3] AI Timeline: Key Events in Artificial Intelligence from 1950-2023. https://www.theainavigator.com/ai-timeline .

[4] The History of Artificial Intelligence from the 1950s to Today. https://www.freecodecamp.org/news/the-history-of-ai/ .

[5] Rise of the Machines: 10 Defining Moments in the History of AI .... https://www.digitaltrends.com/cool-tech/history-of-ai-milestones/ .

[6] AI's Ascendance in Medicine: A Timeline | Cedars-Sinai. https://www.cedars-sinai.org/discoveries/ai-ascendance-in-medicine.html .

[7] An AI revolution is brewing in medicine. What will it look like? - Nature. https://www.nature.com/articles/d41586-023-03302-0 .

[8] Sharing Google's Med-PaLM 2 medical large language model, or LLM .... https://cloud.google.com/blog/topics/healthcare-life-sciences/sharing-google-med-palm-2-medical-large-language-model .

[9] Med-PaLM. https://sites.research.google/med-palm/ .

[10] Scaling Smart Solutions with AI in Health: Unlocking Impact on High .... https://www.weforum.org/reports/scaling-smart-solutions-with-ai-in-health-unlocking-impact-on-high-potential-use-cases .

[11] How will generative AI impact healthcare? | World Economic Forum. https://www.weforum.org/agenda/2023/05/how-will-generative-ai-impact-healthcare/ .

[12] New World Economic Forum Research Identifies Top AI Applications That .... https://www.weforum.org/press/2023/06/new-world-economic-forum-research-identifies-top-ai-applications-that-could-revolutionize-global-healthcare .

[13] AI chatbots are supposed to improve health care. But research says some are perpetuating racism. https://au.news.yahoo.com/ai-chatbots-supposed-improve-health-090703160.html .

[14] How AI is Reshaping Global Health and Bridging Social Gaps. https://www.cryptoglobe.com/latest/2023/10/the-transformative-role-of-ai-in-global-health-and-reducing-social-inequality/ .

[15] 'Everybody was telling me there was nothing wrong' - BBC. https://www.bbc.com/future/article/20180523-how-gender-bias-affects-your-healthcare .

[16] Survey: 34 Million Americans Know Someone Who Died Because They Couldn .... https://www.forbes.com/sites/niallmccarthy/2019/11/14/survey-34-million-americans-know-someone-who-died-because-they-couldnt-afford-medical-treatment-infographic/ .

Andrei Iordache

Founder & Product Mastermind at UPDIVISION ? Elevating UX and Product Strategy since 2010

1 年

This is one area where AI can be very useful, but we should be careful about how reliable its data is.

要查看或添加评论,请登录

社区洞察

其他会员也浏览了