Let's read the text
Required reading time: 3 minutes
Siri, Alexa and Co. allow more autonomy and independence for people with disabilities. The popular language assistants are only one of many possible applications of artificial intelligence in the health care sector: Machine learning made easier diagnostic procedures or the communication between doctors and patients. But there are also disadvantages.
56 million smartspeakers to be sold worldwide in 2018, estimates the global analysis company Canalys. Ovum, a global research and consulting firm, predicts that by 2021 there will be more digital assistants than people in the world. Artificial intelligence is already integrated in washing machines, cars or light switches. In the long run, numerous objects in our environment will understand, respond to, or even act independently of voice commands. But why is this technology so successful?
Less barriers due to language assistants
Language is a natural part of our lives and occupies only a few brain areas. The average person can speak 150 words per minute, but type only 40 words per minute. Voice navigation is therefore faster, less tiring and more intuitive than (almost) any other form of handling.
Another important advantage of Siri, Cortana or Alexa is the accessibility. This eliminates the need to enter text when interacting with a language assistant and thus the need to be able to read and write. A tremendous advance, not only for dyslexics. The eyes are also no longer needed, for example to read through a definition or place an order on the Internet. The wizard reads the written word aloud instead. This allows visually impaired or blind people more convenient access to web content. For paralyzed people, freedom of sight and hands also mean new independence and autonomy. The customer reviews on the Amazon product page for the language assistant "Echo" tell numerous stories about those affected.
If the body's own abilities are limited and the risk of accident increases, Alexa. Google Home and Co. even become life savers. For example, in the event of a fall within one's own four walls with serious injury: a simple voice command informs either the emergency call or relatives within seconds.
Alexa in the nursing home
A pilot project in the Carlsbad by the Sea retirement settlement near San Diego (USA) investigated whether language assistants are useful for seniors. In February 2017, the trial series brought together the Amazon language assistant Alexa and the residents of the community. The aim was to explore the utility of modern technology for older people (80 years and older). The result surprised even the responsible organisation Front Porch Center for Innovation and Wellbeing.
- 75% used their language assistant at least once a day
- 100% had the feeling that Alexa is a relief in everyday life
- 71.43% felt more connected to family, friends and community
The biggest challenge for the focus group was to formulate search queries in such a way that Alexa understands them. In addition, the voice pitch of the intelligent device caused problems for some participants with hearing aids. Amazon is currently checking whether the highs and lows of the smartspeaker can be adjusted manually in future.
Artificial intelligence cannot replace a doctor
The spread of this voice-controlled hardware is evolving synchronously with AI-based health applications. The areas of application go far beyond assistance in everyday life. Machine Learning is the basis for numerous apps that connect doctors with patients quickly and precisely or make complex data more understandable. Artificial intelligence can predict dengue fever epidemics temporally and geographically, detect tumors, help blind people see or diagnose autism in children. The first expert system of this kind was developed by Edward Shortliffe at Stanford in the early 1970s. MYCIN optimized the intake of antibiotics depending on the disease to avoid overdoses.
Despite fundamental progress, AI-based technologies have been slow to establish themselves in medicine. The reason: in healthcare data protection is a particularly critical issue. Patient information in particular is highly sensitive and legally protected for good reasons. Attractive growth figures therefore remain a challenge.
And there are other restrictions: Machines can't express, for what reason they have achieved a result. While computer-generated "suggestions" support a diagnosis, a physician's assessment remains necessary to address the specific context of the patient. AI can therefore accelerate or optimize the diagnostic process, but cannot replace the human expert.
My conclusion:The potential for artificial intelligence and machine learning in healthcare is immense. The new technologies remove barriers and increase the accessibility and efficiency of an overburdened health care system. The challenges lie in data protection and the sensible networking with human experts.
Cover picture: Photo by Michael Moeller, EyeEm