Artificial Intelligence Is Essential For Our Ecosystem

Allow me to start this article with a slightly provocative statement. If you believe that you and your organisation can have a future without Artificial Intelligence (AI), then you might be similar to those who thought that there is a future without the Internet in 1996 and without Electricity in 1882.

The main points of this article is that:

- AI is the most disruptive class of technologies over the next 10 years.

- AI is an essential tech layer to create high quality & competitive services and products

NAII Article Series #2 - Artificial Intelligence insights from the Nordic Artificial Intelligence Institute (www.nordicAIinstitute.com).

by Christian Guttmann Executive Director of the Nordic Artificial Intelligence Institute

Just as the Internet and Electricity were becoming essential technologies at these times, Artificial Intelligence is becoming essential to our personal lives, business organisations and society at large. Without AI, your organisation will not continue to do competitive business and create meaningful products and services at scale. Similar to having an email and a website for improved communication today, you will use (directly or indirectly) AI when driving your car and receiving better health care.

Figure 1: “Artificial Intelligence: the most Disruptive Class of Technologies” -- Source Gartner (July 2017)

In line with our own observations, Figure 1 shows a few remarkable insights on the AI development by Gartner (July 2017). According to Mike J. Walker (Research Director, Gartner):

"Artificial intelligence technologies will be the most disruptive class of technologies over the next 10 years due to radical computational power, near-endless amounts of data and so on... these will enable organizations with AI technologies to solve problems that no one has ever encountered previously."

A substantial projection by Gartner is that Artificial General Intelligence (AGI) is identified as one of the most important technologies over the coming decades (Figure 1: AGI is the third yellow triangle from the left). This is interesting due to AGI (also referred to as Strong AI) being one of the most ambitious goals of Artificial Intelligence. An AGI system is a combined system capable of performing many complex tasks normally performed by narrow AIs. A narrow AI system is very good in performing one task (e.g. driving a car or detecting a tumor in an MRI), but they are not integrated or connected. AGI combines many of the narrow AI technologies on the right side of Figure 1, including Deep Learning, Machine Learning, Augmented Data Discoveries, and Smart robots. In short, AGI is a big idea! 

To exemplify AGI in a more practical context: you will have one AGI that can influences many aspects in your life. Your AGI will self-drive your car, monitor your health while you are in your car, and also at home and work, the AGI will have a dialogue with you, and it can cook, be aware what is in the frige, and clean up the kitchen. In the background the AGI will work with those entities that you allow it to work with, e.g. your medical facility, your bank, and your retailer. Please read here about how narrow AIs are used in health care.

Figure 2 shows several layers of technologies that have been essential to create current and future services and products. Artificial Intelligence is one of these essential business layers that more recently has been added to the stack.

Figure 2: Artificial Intelligence is an essential tech layer for developing Services and Products, just as essential as Smart Phones, Internet, Computers, and Electricity (there are more layers, but omitted here for sake of clarity).

As you are looking at Figure 2, ask yourself this question: What type of services and products could be developed today by removing any of the essential technology layers?

Only a few products and services will be advanced and you will also not be broadly roll them out. Even the development of the simplest of products often require the use of many (indirect and direct) essential technologies -- consider a very simply example: producing and selling a loaf of bread relies on advanced communication (e.g. emails and calls to suppliers and transport agencies) and basic electricity (e.g. bread baking and lights in the bakery shop). AI is increasingly used across the value chain for any product and service.

Here is a short scenario that demonstrates the readiness of a country or organisation to ensure AI is used as effective as possible with a benefit for the ecosystem and society.

Think of two countries/organisations -- country/organisation A and country/organisation B.

Country/Organisation A has a clear AI strategy: a well thought out AI policy, a strong AI infrastructure, and competency roadmap across all industry sectors. Country A's strategy aims at longer-term transformational impacts of AI on society and the world, identifies strategic priorities for near-term and long-term support of AI that address technical and societal challenges. Country A can also ensure that products and services meet international AI quality standards (such as ISOs) and international AI ethical guidelines.

Country/Organisation B has fallen behind on creating an AI strategy: it has not provided a framework for industry and it has set aside no sustainable budget funding for education and innovation. Which country is more likely to create and ensure higher living standards, longer healthier lives, and competitive and higher quality services and products?

A similar comparison can be made for companies and other organisations. Country A will of course do better in such a scenario, just as it would do better compared to a country that does not have a good Electricity or Internet infrastructure.

Here are two considerations that should be met to ensure a sustainable ecosystem and continue the creation of high quality services and products. Governments and government authorities make Artificial Intelligence a strategic priority across all sectors and create a clear AI strategy. Business can then plan effectively. Companies in key industry sectors create strong a AI leadership by assembling the brightest minds in AI, collaborate with leading research institutions and develop a strong internal AI roadmap that is executable. One important point to make: AI talent is extremely hard to come by.

What an AI strategy, priorities and leadership would entail will be discussed in a next article in this AI series. Stay informed by simply following Christian Guttmann on Linkedin and Twitter, and follow the Nordic Artificial Intelligence Institute. 

Christian Guttmann (PhD) is the executive director of the Nordic Artificial Intelligence Institute, Associate Professor (Adj.) at the University of New South Wales, and Research Fellow (Adj.) at the Karolinska Institute. Christian is also in the organisation of the joint conference on Artificial Intelligence (IJCAI). With over 20 years of experience in leading international projects on AI in Health Care and Medicine, his expertise is in the intersection of research, innovation and industrial/commercial application of AI. He has a PhD and degrees in Artificial Intelligence and Psychology.

Introduction Map of Artificial Intelligence in Health Care and Medicine

Article series from the Nordic Artificial Intelligence Institute (www.nordicAIinstitute.com).

Main Points:

• Introduce a map, overview and guide describing significant areas of AI and Health Care/Medicine

• Explain briefly scientific disciplines contributing to Artificial Intelligence

• Discuss practical AI technology areas that transform Health Care/Medicine

Artificial Intelligence (AI) is a term coined by John McCarthy in 1956. This also marked the beginning of many substantial scientific and industrial AI activities (including the IJCAI conference series - the world's premier scientific conference on AI). But more so, AI has become a dominant transformational driver in the health industry and the public sector. This AI transformation comes at a good time - health care systems face huge challenges in continuing to deliver quality of care - one reason is the increase of chronic diseases.

The good news is that AI offers substantial benefits to the health care and medical field. AI is becoming increasingly accurate and effective in performing a broad range of complex health care related tasks (e.g. recognizing a malignant tumor on MRIs). Such AI driven performance can then be scaled up, improve health outcomes and save many lives. As a result, AI helps clinicians, patients and many health stakeholders to make faster, better, and cheaper decisions, at scale. Figure 1 shows a map of some of the primary areas of AI and Health Care/Medicine.

Figure 1: An introduction map, overview and guide of Artificial Intelligence in Health Care and Medicine.

The intersection of AI and Health Care & Medicine is a rich field of science, technology and innovation. As seen on the left in Figure 1, scientific areas that contributed to the field of Artificial Intelligence are Philosophy, Neuroscience, Health, Medicine, Mathematics, Computer Science, Economy and Engineering. For example, philosophy supports our understanding of key concepts in AI such as the philosophy of mind and "free will". Neuroscience and AI have been intertwined for decades, resulting in concepts such as Artificial Neural Networks. There are of course many more scientific disciplines than those listed in Figure 1, and each contributes in different ways to the progression of AI.

AI has brought forward many advanced technologies, and is used in many companies and organisations (see e.g. the recent Health 2.0 conference in California). Figure 1 names just a few in the context of the health care and medical field, discussed here in more detail.

Robotics Applications in health include robotic surgeons (robots perform basic surgery autonomously and learn from human surgeons), medical drones (delivering blood and defibrillators on demand), robotic nurses and companion robots -- medical cobots.

Machine Perception AI technologies that use their "sensors" similar to the way humans use their senses. This may include the recognition of malignant tumors in MRIs, and "smelling" if you are sick based on a person's odorprint.

Natural Language Processing Related areas are computational linguistics, natural language understanding, and multi modal interactions (including e.g. facial expression and intonation). These have resulted in application such as chatbots used at the NHS in the United Kingdom to answer medical questions, understanding unstructured medical notes into standardised electronic health records, and suggest treatment options based on understanding millions of scientific medical publications. None of these tasks could be done by health care professionals at scale.

Machine Learning: Due to a massive increase of data about the human condition (e.g. physiological and behavioural information) and trying to make use of this data, machine learning has received a lot of attention in recent years (e.g. related areas are analytics, data science, data mining, approaches which still need a human in the loop). Machine Learning applications are vast, e.g. classifying DNA sequences and medical diagnosis. This area also comes with fundamental ethical questions not currently addressed, e.g. digitising cultural prejudices.

Multi Agent Systems: Coordination and collaboration of multiple actors is critical in many health care settings. An AI system supports care teams that often consists of multiple health care providers, carers, and the patient themselves. The AI models used in Multi Agent systems allow for smooth coordination among several entities, such as individuals, clinics, and authorities. This area also discusses the team interaction of mixed human-AI teams.

AI is much broader and deeper than explained in this article. Some countries and organisations are well ahead in the era of AI, while others are just at the beginning. It is clear that countries and organisations placing AI clearly on their roadmap, have a strong AI strategy, and are able to execute on it, will have a better position in the future. We soon publish an article on which organisations are ahead in the AI era -- stay tuned.

We continue this article series on Artificial Intelligence. Stay informed by simply following Christian Guttmann on Linkedin and Twitter, and follow the Nordic Artificial Intelligence Institute. With great dedication, Christian has been leading themes and projects in each of the above AI areas over the last 20 years. If you would like to discuss this topic, please reach out and leave your comments below.

Christian Guttmann (PhD) is the executive director of the Nordic Artificial Intelligence Institute, CEO of HealthiHabits AB, Associate Professor (Adj.) at the University of New South Wales, and Research Fellow (Adj.) at the Karolinska Institute. With over 20 years of experience in leading international projects on Artificial Intelligence (AI) in Health Care and Medicine, his expertise is in the intersection of research, innovation and industrial/commercial application of AI. He has a PhD in Artificial Intelligence, and several degrees in Artificial Intelligence and Psychology.