An increasing number of businesses and organizations are ramping up AI deployment, giving a boost to mass adoption of AI across all sectors and industries in a time of Covid-19 pandemic. AI is becoming a module essential to the supply of products/services and everyday running of numerous enterprises and is deeply entrenched into the development of industries. All AI-related technologies are also evolving into strategic resources of vital importance to the revolution of social productive forces, and are expediting the digitalization of all sectors, including high-tech manufacturing, medical and healthcare, finance and consumption. Among them, the medical and healthcare sector is a booming sector receiving lots of attention and investment in recent years. This article is intended to explore the phenomenon of healthcare being preferred by AI and the reasons behind from the perspectives of AI adoption by industry, size of private investment and best practices in market segments. The author wishes to discuss the topic with other peers in the industry and is open to advice.
AI Adoption by Industry
To answer the question why the AI technologies get more attention and commercial use in the medical and healthcare sector, we, first of all, need to see the big picture of the technologies involved across the AI industrial chain and where the industry is going.
The AI industrial chain is composed of the basic layer, technology layer and application layer. The basic layer mainly provides basic support, such as data resources, hardware and computing power platform.To be more specific, there are AI chips, sensors, big data, and cloud computing. Among them, big data and sensors are primarily used in data acquisition, and chips and cloud computing in data analysis and computing.The technology layer is the key to the development of the AI industry, including universal technologies, AI technical framework and algorithm models.This layer is about developing all sorts of applied technologies by means of big data mining & processing and machine learning modeling, and then solving real-world problems. Some of those typical technologies are computer vision, natural language processing, speech recognition, machine learning and deep learning.The application layer addresses the commercialized applications of AI in varied fields,inclusive of but not limited to security, finance, retail, education, culture & amusements, marketing, agriculture, manufacturing, transportation and services.
AI Industry Mix and Typical Companies |
Products and Solutions Layer |
Security(HIKVISION),retail(AliCloud, Tencent Cloud, Baidu Cloud),medical & healthcare(SenseTime, Infervision, Shukun),urban mobility(Didi, Apollo),mobile terminals(Huawei, SenseTime),client services(Baidu, Xiao-i Robot),finance(Ping An Technology, Cloudwalk),education(Squirrel AI),culture & amusements(Moviebook, iQiyi),marketing(Moviebook, Youku),agriculture(AliCloud, Baidu), andmanufacturing(AliCloud, Baidu) |
Technology Layer |
Computer vision(SenseTime, ArcSoft, Baidu, Cloudwalk),smart speech recognition(iFLYTEK, Unisound),natural language processing(Baidu, Unisound, Mininglamp Technology),knowledge graph(Mininglamp Technology, Baidu), andmachine learning(Google, Baidu, 4Paradigm, Microsoft) |
Basic Layer |
Chips(NVIDIA, Cambricon),platforms(ArcSoft, iFLYTEK),open-source framework(Baidu, Google, Microsoft),sensors(HIKVISION),cloud services(Alibaba, Baidu, Tencent, Huawei), andbig data services(Cloudwalk, Mininglamp Technology) |
With knowledge of the classification of AI technologies and its applications by industry, we may look into the adoption of the AI technologies in the medical and healthcare sector. The “Adoption Rate by Industry” is defined as the percentage of those parts of a given industry that have adopted the AI technologies (e.g., machine learning, computer vision and natural language processing), according to The State of AI in 2021 by McKinsey.
The chart below shows AI adoption by industry and function in 2021. The adoption rate is the highest in product and/or service development in the high-tech /telecom industry (45%).The AI adoption rate in the healthcare systems/pharma and medical products industry is not the highest but it is up to 29% in product and/or service development, that is pharmaceuticals R&D. Moreover, all the rest parts of the entire healthcare industrial chain have adopted the AI technologies to varying degrees.AI has a high adoption rate in the healthcare sector, particularly in pharmaceuticalsR&D, becauseit has unrivaled advantages over human beings in terms of its ability to gather and analyze in depth an enormous amount of data.Meanwhile, given limited medical resources globally,the application of the AI technologies to different parts of the healthcare chain can improve the efficiency and effectiveness of diagnosis and treatment across the entire healthcare system.Different segments of these applications and practices will be discussed more elaborately later.
(AI Adoption by Industry and Function, 2021, Chart: 2022 AI Index Report)
Size of Private Investment
The preference for “AI+Healthcare” is also evidenced by a sizable amount of private investment flowing into the healthcare sector globally over the last five years. This has much to do with the global capital’s craze for AI in general.In 2021, global private investment in AI totaled USD 93.5 billion, more than double the amount in 2020. Nevertheless,the concentration of investment increased.The number of AI firms receiving new investment dropped to 746 in 2021, from 1,051 in 2019 and 762 in 2020. At the same time,the amount of fund-raising per deal climbed up.In 2021, there were 15 fund-raising deals exceeding USD 500 million, compared to only 4 in 2020. This also suggests the industry is becoming more purposeful and rational instead of going wild with extensive growth.
(Private Investment in AI 2021, Chart: 2022 AI Index Report)
Amount of Fund-raising |
2020 |
2021 |
Total |
>USD 1 billion |
3 |
5 |
8 |
USD 0.5-1 billion |
1 |
10 |
11 |
USD 100-500 million |
93 |
235 |
328 |
USD 50-100 million |
85 |
194 |
279 |
<USD 50 million |
2102 |
2120 |
4222 |
Not disclosed |
354 |
395 |
749 |
Total |
2638 |
2959 |
5597 |
The biggest amount of private investment in AI in 2021 went into data management, processing, cloud (about USD 12.2 billion), 2.6 times the amount invested in the same field in 2020 (about USD 4.69 billion), according to data disclosed by Stanford.The second largest investment was received by medical and healthcare (USD 11.29 billion), followed by Fintech (USD 10.26 billion), AV (USD 8.09 billion) and semiconductor (USD 6 billion).Five-yearsum data (2017-21) shows medical and healthcare raked in the largest amount of private investment globally on a cumulative basis (USD 28.9 billion),followed by data management, processing, cloud (USD 26.9 billion), Fintech (USD 24.9 billion) and retail (USD 21.95 billion).
(Left: Private Investment in AI by Focus Area, 2020 vs. 2021, Right: Private Investment in AI By Focus Area, 2017-21 (Sum), Chart: 2022 AI Index Report)
Medical and healthcare was able to attract such large-sized funding, perhaps because of the following reasons. First,the data sources of the medical and healthcare industry were wide-ranging and of enormous magnitude, but generally had poor availability. The fast development of basic AI technologies in recent years has led to a giant leap in algorithm and computing power. Hence, AI can better empower the medical and healthcare industry and ramp up mass commercialization. Second, scarcity of medical resources is a global issue.Digitalization of the medical and healthcare industry can benefit the human race.Therefore, driving the combination of AI and healthcare means a lot. Third,the medical and healthcare industry provides vastly different application scenarios to different AI technologies.There are numerous market segments with unlimited possibilities and bright prospects. As the AI technologies mature, it is possible to realize differentiated application and implementation of these technologies under various scenarios.
Best Practices in Market Segments
In the medical and healthcare sector, AI is mainly applied to new drug R&D, adjuvant diagnosis, health management, hospital management and patient services. The table below lists AI application scenarios, and AI technologies used in the said six aspects, as well as typical players in each aspect and their best practices:
pplication |
Scenarios |
Key AI Technologies |
Best Practices of Chinese/Foreign Companies |
New drug R&D |
Used to carry out crystal prediction, pharmacovigilance and target discovery in new drug development |
Machine learning |
Insilico Medicine: ·Founded in the USA in 2014, it is committed to developing alternatives to animal testing employed in pharma R&D; its AI engine is able to analyze how drugs act on cells and their potential side effects, and serves a number of top pharmaceuticals and cosmetics companies; ·It has invented an AI system applicable to drug development, which is able to create brand-new molecules within 21 days at the cost of only USD 150,000. It beats the about 95% failure rate in target discovery, tackling one of the biggest bottlenecks in drug discovery. |
Adjuvantdiagnosis |
Improve the efficiency and effectiveness of diagnosis with the help of surgical robot, DNA testing, medical imaging, CDSS (Clinical Decision Support System) and adjuvant therapy |
Machine learning, computer vision, speech recognition |
Intuitive Surgical: ·Founded in the USA in 1995, it mainly studies robotic-assisted, minimally invasive surgery, with its renowned product, da Vinci® surgical system, applied to prostatectomy, repair of cardiac valves and gynecological and obstetric surgery; ·There were more than 6,700 active da Vinci surgical systems installed globally as of the end of 2021, and more than 1.5 million procedures were performed with it in 2021. And clinical data indicates robotic-assisted intestinal cancer surgery has salient advantages. |
|
|
|
|
Health management |
Use smart wearable devices and other smart terminal devices to monitor health data, and produce health management reports, mainly serving the health checkup market |
Data analysis, computer vision, speech recognition, knowledge graph |
ClouDr: ·Founded in China in 2014, it is a one-stop chronic disease management and intelligent healthcare platform; ·Its hospital SaaS inquiry system covers more than 1,800 hospitals and is able to better assist these hospitals in comprehensive, timely, dynamic chronic disease management for patients, with a city penetration rate of 88.2% and mainstream hospital deployment rate of 20% across China. |
Hospital management |
Electronic medical records (EMR), medical supplies management, financial management, manning, and traffic monitoring |
Database management, computer vision, natural language processing, speech recognition |
Epic Systems: ·Founded in the USA in 1979, it is a medical and healthcare software company; ·It mainly develops and sells proprietary EMR. In 2022, the hospitals using its software owned the medical records of 78% of all patients in the USA and over 3% of all patients globally. Goodwill: ·Founded in China in 2006, it is among the first companies in China involved in medical information system software R&D and industrialization; ·It claimed the top spot in China’s EMR market ranking for seven years in a row but is not very profitable. |
Patient services |
Intelligent inquiry and guidance, hospital navigation, patient visit |
Speech recognition, machine learning, knowledge graph |
Envive: ·Founded in Shanghai in 2016, it has developed and launched multi-disciplinary inquiry and triage systems, covering more than 4,000 diseases in 12 major medical departments; ·Key founders have Stanford University and Silicon Valley background and its product took second place in WAIC 2020; ·Its inquiry system enables physicians to browse through patients’ vital signs, main needs, current medical history and related medical records within 3 to 5 seconds upon their arrival at the consulting room. |
Others |
Aesthetic medicine, medical payment, knowledge management |
Machine vision, deep learning algorithm, medical knowledge graph |
Allergan Aesthetics: ·Founded in Ireland in 1950, and acquired by AbbVie in 2019, it mainly deals in ophthalmologic and medical aesthetic products; ·Empowered by Microsoft’s AI technologies, it is able to provide individual consumers with intelligent diagnosis, medical aesthetic effect simulation and other customized services. |
Generally speaking, the “AI + Healthcare” mode is much-coveted with high adoption by the industry and stunning performance in the capital market, in the eye of the author, for the following reasons.First, there is an enormous build-up of data in this field and AI provides unrivalled computing power. Second, the digitalization of the healthcare sector brings along universal benefit and value and is worth the focused effort and heavy investment. Third, the AI technologies have reached a higher level of maturity and variety in recent years.There has been a roll-out of plentiful research results in machine vision, natural speech recognition and processing, data mining and machine deep learning. Hence, the AI technologies are more widely and deeply applied to differentiated scenarios in the healthcare sector. As a whole, the degree of AI commercialization varies from segment to segment of the healthcare sector. More applications are seen in drug R&D and adjuvant therapy, while commercialization in health management and other segments could be quicker.
In the future, in the process of intelligence-driven development of the healthcare sector, priority should be given to reducing the cost of data, and improving the awareness and capability of data security and protection. Ethical issues coming with AI applications should be considered. Meanwhile, it is important to explore sustainable business models, heighten the willingness and capability to pay of both businesses and consumers, and discover diversified and sustainable revenue streams.
References:
1，2021 China’s AI Industry Report (IV), iResearch.
2，The-state-of-AI-in-2021, McKinsey, December 2021.
3，Daniel Zhang, Nestor Maslej, Erik Brynjolfsson, John Etchemendy, Terah Lyons, James Manyika, Helen Ngo, Juan Carlos Niebles, Michael Sellitto, Ellie Sakhaee, Yoav Shoham, Jack Clark, and Raymond Perrault, “The AI Index 2022 Annual Report,” AI Index Steering Committee, Stanford Institute for Human-Centered AI, Stanford University, March 2022.
4，A Panoramic White Paper of The Digital Economy, Analysys
5，IDC official website
6，Guangwen Guangzhi: AI Report-Volume I
7，CICC: AI Applied to Boost Improved Efficiency and Resource Sharing in Healthcare
