In the heart of Berlin, there is a futuristic building – a hybrid of a museum and a lab. Inside, the visitors are confronted with a challenging question: How do we want to live in the future? I visited the FUTURIUM to explore the AI in the healthcare of tomorrow. Watch the video and read an interview.
How will AI change healthcare in the future?
Today, scientist and doctors already have access to a wealth of information: new research findings and patient healthcare data is growing. AI can help health professionals to deal with this flood of data – faster and more precisely.
AI can show correlations that might be invisible to humans. With the guidance of AI systems, new therapies can be developed. Scientists could find out more about the individual characteristics of illnesses and develop personalized medication. Intelligent expert systems could also assist doctors in diagnosing certain diseases or planning the most effective treatment. In the best case, physicians could, therefore, have more time for personal interaction with patients.
Also, intelligent machines can work directly with our minds. They can discover patterns in the human brain and use them to operate a device, for instance, a prosthesis. The connection between brain and machine is also expected to help paralyzed people communicate.
Could you give some examples of AI implementation today?
Expert systems can learn to detect diseases by comparing data and images of healthy and unhealthy patients. A full-body scanner, for example, is used for skin cancer prevention: the computer program can automatically and thoroughly scan the human skin surface and detect differences. Brain-computer interfaces can help people suffering from locked-in syndrome. Those patients are unable to move and speak. Using only the power of their minds, some of the patients can write a text on a computer.
What are the most critical questions regarding AI in healthcare and AI in general?
One crucial question is: Will medical practice soon be wholly taken over by computers? And how much decision-making power should AI have in the future?
We need to be aware of the fact that usually, a visit to the doctor is a complex circumstance. Computers – to be able to replace doctors – would need to incorporate many more factors into their decision-making process, for example, an analysis of the patient’s behavior. Not to mention other factors, like the experience and empathy of the medical professionals that play an essential role.
Another question we need to discuss is how we would want to deal with the data that is produced by applying AI in healthcare and how much we want to rely on the decision made by machines. Doctors but also patients – on a more fundamental level – must understand how digital technology is used and how their data is stored and protected. This sensitive, personal data that includes health records must be well-guarded and processed transparently. Otherwise, it could potentially be used to judge or classify people in an unethical way.
And, we have to be aware that humans can no longer reconsider some automated, AI-driven decisions. So, errors made by algorithms with potentially serious consequences can occur if physicians rely entirely on the data supplied by the computer.
Which aspects of the AI have been exhibited in Futurium?
Digitalization, AI, and Big data are topics that play an essential role in many parts of the exhibition. For example, when we discuss the future of work or how we manage life in big cities. AI can also help to monitor and improve air quality in the city, which affects our health, too.
Healthcare-related AI is a topic we discuss when we talk about the increasing amount of health data in science and how this could be used to understand diseases better. One fascinating aspect here is: What role should AI play in doctors’ decisions? And how can health data be protected?
We also talk about the interaction with healthcare-related machines like high-tech prostheses, brain-computer interfaces, or care robots. On the one hand, we show solutions that will directly affect our lives in the future and ask intriguing questions: How will diseases be treated in the future? Am I still going to see a doctor?
On the other hand, we present the significant challenges of today, discuss controversial issues from various perspectives, and encourage our visitors to think about how we embrace technologies. There are so many topics to be addressed: Who will take care of us when we age? Which ethical rules might algorithms need?
Can we predict the future of AI if we really don’t understand how it will develop?
Thinking about ‘artificial intelligence’ means that we conceptualize it in terms of human intelligence. Right now, researchers are still trying to fully understand how intelligence as a concept works to incorporate it into machines. Many disciplines need to work together: neuroscientists, computer scientists, psychologists. AI could be used to understand better how human intelligence works.
How will life with algorithms look like?
We will be more connected in the future – in our workplaces and neighborhoods, with the digital devices that we always have with us. Algorithms are already today processing vast streams of data, and we don’t realize that. They analyze how the environment is changing, how we behave, and what we need. As a result, we gain new knowledge that can be accessed at any time and might improve how we work and live.
But with the comfort and security that automation delivers, we also give up responsibility and our own autonomy. Because whoever has control of the data and the tools necessary for processing, will decisively shape the merging physical and virtual worlds.
How to ensure ethically developed algorithms?
The more powerful algorithms that shape our society, the more urgently we need general principles for the use and control of AI systems. Projects like Algo.rules (Rules for the Design of Algorithmic Systems) which we present in the exhibition, develop guidelines for people who have a significant influence on the creation, development, programming, use, and effects of algorithmic systems.
These guidelines want to make programmers aware of questions considering their work: What will a particular algorithm be used for? Is the training data biased? How to ensure diversity of all types among people responsible for the development of algorithms?
Ultimately, solutions must be developed, whether they are technical or political, to ensure that these systems are used to serve society as a whole.
Some say that AI could be the biggest threat to humanity. Do you agree?
While most of the current AI applications are used to impact humankind positively, any powerful tool can also be used for harmful purposes. This is why we now need to define ethical guidelines to ensure that we are certain about how and for what AI is applied.
And we need to make sure that everyone understands better what AI can and can’t do. Machines can be precise at a specific task, but they are unable to complete complex, spontaneous or creative tasks. So, we need to educate people on the possibilities of AI without scaring them off with unrealistic scenarios that do not reflect the state of research. There is always a human being who writes a program and defines how it is used.
Special thanks to Jasmin Minges (Research Associate) and Antja Karoli (Trainee) for hosting me in the FUTURIUM, answering all the questions, and assisting in making a video. The recordings, photoshoots, and interviews couldn’t be possible without the big help of Monique Luckas (The Lead of the Communication Department).
I have a small favour to ask…
This content is free of charge. This website is free of commercials. Please support aboutDigitalHealth.com (€1+). It only takes a minute. Thank you!