Artificial intelligence (AI), i.e. self-learning computers, already accompany people in many walks of life today. Language assistants help with buying groceries and play music via voice command. Household robots make daily chores easier. Cars are beginning to park themselves.
The area of digital language assistants in particular is moving fast: People don’t just communicate with a computer operating system anymore, like Windows or iOS. Digital companions such as Siri and Alexa are on the rise. Artificial intelligence can efficiently and quickly perform tasks that people haven’t mastered or which are very difficult. It’s a handy tool that can facilitate people’s everyday lives by taking care of routine tasks. And at the same time, it offers access to extensive knowledge without great effort.
But artificial intelligence’s success is dependent on a task only having one unique solution. Because intelligent computer programs training to become intelligent. They have to know that the result they arrive at is the right one. Intelligent systems are used in medicine, for example, for image-based diagnostic procedures. AI systems have been trained with large image databases, and they’ve been able to predict malignancy of tumors with success. Other self-learning systems can find out whether certain pains are symptoms of a serious condition. Many health professionals hope that AI can deliver quick and precise results. Doctors can then use the time saved to devote themselves more intensely to their patients.
It’s another issue altogether when a question or task has no clear answer. Can machines drive cars better, for instance? If the answer is limited to compliance with rules in road traffic and accident-free driving, an autopilot could prove much better. Machines’ choices, however, are very different from human drivers’ split-second decisions, which are informed by emotional and moral principles and not just logic. This leads to ethical questions, which so far have not been resolved.
It is not surprising that many people are critical of self-learning computers. “In short, success in creating effective AI could be the biggest event in the history of our civilization, or the worst,” said the British astrophysicist Stephen Hawking. In his opinion computers could eventually emulate and even surpass human intelligence altogether. It could develop its own will.
Also hidden behind the results, the AI technologies deliver big data analytics. Nearly every user action in the digital space is associated with the collection and processing of data. Every day huge amounts of data are generated that are used by manufacturers, service providers, banks, insurance companies, employers, scientists and investigating authorities for their own purposes.
Criticism of artificial intelligence also comes from the fact that how algorithms work is incomprehensible for many people. Even data scientists don’t always know how self-learning computers come up with their solution strategies and what they will do next.
That worries people: Will we ever understand how an artificial intelligence makes its decisions? “Alice” and “Bob”, the two networks from before, are proof that AI can keep a secret. Hence, it’s paramount that data scientists not only create intelligent systems but also build in the needed transparency into how they operate.