Is IBM's cognitive computing overwhelmed

Cognitive Computing - Cogs : Talented machines

In Douglas Adam's sci-fi novel The Hitchhiker's Guide to the Galaxy, it took seven and a half million years. The question about life, the universe and everything else seemed so big that only a computer could answer it. The new computer should be smarter than its inventor. It was built as big as a small town. His name: Deep Thought. But on the day when he finally finished the calculation, he just spat out one number: 42. Nobody could do anything with it.

Inspired by this episode, the American computer company IBM baptized a chess computer with the same name in the late 1980s: Deep Thought. A good two decades later, IBM is preparing to build a new supercomputer. The company wants to usher in nothing less than a new era in computer history. The name of the supercomputer: Watson, named after IBM founder Thomas J. Watson. The technology behind it is called cognitive computing. In 2011, Watson was still as big as a living room, he won against two human challengers in "Jeopardy!", A more elegant variant of "Who Wants to Be a Millionaire?" Today, just four years later, Watson has shrunk to the size of a stereo system. Soon it could fit in a pocket.

So far, the old saying about the computer was considered a stupid machine that only does what a person tells it to do. Cognitive computing is changing the paradigms. Cognitive computers, or cogs for short, are capable of learning, store information and continue to develop. Cogs can be imagined as gifted children who can absorb everything and process it at a speed that humans can only marvel at. The artificial intelligence that researchers have been working on for years is about to be launched.

Smarter than their creators

Machines that are intelligent and, as in the fictional example of “Deep Thought”, possibly more intelligent than their creators, have been inspiring and disturbing humanity for decades. For some, cognitive computers - also in IBM's PR parlance - promise to set the course at a time when the flood of information overwhelms many. You should get the best result from all the data. For others, they are a further step towards a superintelligence, which - so the concern - could exceed human intelligence and possibly be dangerous to us. For the British physicist Stephen Hawking, for example, who recently warned that cognitive computers could mean the end of mankind. Hawking is also one of the signatories of an open letter pleading for artificial intelligence to be used only for good purposes, to ensure that they are aware of the danger. In addition to Hawking, a number of researchers and engineers have signed this letter - including some who worked on the development of Watson.

Chris Biemann, computer science professor at the TU Darmstadt, does not believe in a scenario in which intelligent computers first outperform us and then want to get down to business. “Why should you do that?” He asks. It is not the technology as such that is dangerous, but how we humans deal with it. According to Biemann, it takes an increased sense of responsibility. “It's similar to cars,” he says. They also got faster and faster. And a car that drives 100 kilometers an hour is more dangerous than one that only drives 20 kilometers an hour.

That we stop thinking for ourselves

IBM is not the only computer manufacturer - with Watson and the “Jeopardy!” Episode, however, the company is considered a pioneer among the public. In fact, many companies are working on cogs. That is understandable, the possible uses are diverse. In diagnostics, for example. "A doctor, no matter how many cases he has experienced in practice or in theory, will never be able to access as much data as an artificial intelligence could," says Manuela Lenzen. She is a research assistant at the Center for Interdisciplinary Research at Bielefeld University. She also works as a freelance science journalist specializing in philosophy and cognitive science.

For Lenzen, the advantage of the Cogs is their specialization. She sees the danger in the fact that artificial intelligence affects all areas of daily life and that we rely too much on the super-intelligent machine. Lenzen even fears that we will stop thinking for ourselves and no longer dare to make our own decisions without first consulting Cogs. "If everyone can access a kind of giant Wikipedia that also knows the right answer to ethical and moral questions and acts accordingly, then I dread it."

For Biemann, this danger does not come from the Cogs. “The interesting thing about the technology is that it includes the reasons and sources,” he says. It remains to be seen whether anyone will bother to check these sources or whether they would prefer to believe in the hyperintelligent computer. What is clear, however, is that scenarios like this, as inconceivable as they may be for some, can nonetheless become reality. Who could have imagined twenty years ago that friends would meet in a bar and then stare at their miniature pocket computers all evening and communicate with other people who were not there?

When the computer is blindly trusted

It becomes critical for Lenzen when people can no longer understand how something happens when the computer is trusted blindly. “If cognitive computers can help to improve traffic light switching, for example, I think that's good and legitimate. I feel uncomfortable when the new technology is used to answer questions that concern us as human beings. ”For example, when a person's life-sustaining machines are switched off.

Biemann does not rule out potential dangers, such as the possibility of manipulation. “Even today there is the sentence: Everything that cannot be found with Google does not exist. The things do exist, of course, but they are much more difficult to find. ”The question remains, and Lenzen and Biemann agree: Ultimately, we are all, society and politics, responsible for what we do with the new technologies .

In any case, the story of Douglas Adams doesn't end well. Because the answer from "Deep Thought" does not satisfy anyone, you first want to agree on the right question. Again, a specially created supercomputer calculates several million years. In the end, however, five minutes before the goal, the earth is destroyed by hostile aliens. The scenario sounds absurd, of course. The supercomputer, however, will be available soon.

Now new: We give you 4 weeks of Tagesspiegel Plus! To home page