We Humans and the Intelligent Machines. Jörg Dräger

Чтение книги онлайн.

СКАЧАТЬ or stressing them. Bettermarks is now available in every classroom in Uruguay; in Germany in can be found in just a few hundred schools.

      Educational systems like the one in Germany, however, would be well advised to abandon the illusion of homogeneity. People are different and they learn differently, too. Even if everyone had to achieve the same goal, their path, style and pace to get there would be very different. Schools too rarely take this into account and often still focus on the average learner. “You’re 12, it’s autumn, so it’s time to do fractions,” says journalist Jürgen Schaefer.¹⁰

      In fact, the achievement range in mathematics for 15-year-olds spans several years, even at Germany’s more challenging secondary schools. Nevertheless, individual support and personalized learning plans are still rare. The result is that many students hate math. Whether fractions, algebra or calculus – once you have missed the boat, you rarely catch up again. Traditional teaching can hardly make up for this: Repeating the material until everyone has understood it would not serve the best students in the class; simply continuing on would leave too many behind. Teachers like Elke Stuthmann want to solve this problem and support each of their students individually by providing digital support. Even then, not everyone will love mathematics, but at least it will prevent a lot of frustration and many hours of expensive after-school tutoring.

       Teach to One

      What is just starting in Hamburg has gone full circle in New York.¹¹ That is where the algorithmic personalization of learning can be found, hidden away behind a brick facade from the 1930s. From the outside, little suggests that the David A. Boody Intermediate School in Sheepshead Bay is already home to the math lessons of the future.

      The ground floor reflects the American norm, with scuffed floors and fluorescent lighting in the hallways, which lead to classrooms on the left and right. On the first floor, however, a new world begins. There is just one large room where some 90 students from three grades can sit together. The building’s former classrooms have disappeared, replaced by a few work areas, separated from each other by dividers and containing chairs of different colors. Electronic monitors hang on the walls. In the morning, the 11- to 14-year-old students stand in front of them looking for their names and their individually assigned learning program for the day. Gazing up, they resemble passengers at an airport trying to identify the gate they are departing from.

      The difference, however, to mass transportation could not be greater. All that matters at Boody are individuals and their individual needs. Simply put, the school has revolutionized mathematics teaching. Nine years ago, the traditional classroom model used for centuries was abandoned, the textbooks mothballed. In turn, the lesson plan was overhauled and digitized. The students now learn in groups or on their own, sometimes with a teacher as a tutor, sometimes with an educational game or explanatory video on their computer. Some eighth-graders still struggle with fractions, some sixth-graders can already handle material meant for their older peers. Everyone can learn as fast or as slowly as they are able. Every day, the students take a brief online test to see if they have reached their individual goal at the end of the lesson.

      Algorithms in Manhattan evaluate the results. That is where New Classrooms is based. The non-profit organization has developed the new teaching concept used at David A. Boody School and almost 40 other institutions in the US, as a response to the enormous diversity among students. Teach to One is its programmatic name. Each day, New Classrooms collects data to see what students have learned. Which tasks have they mastered, which not? How long did it take? How often are videos watched? Are games preferred over text? The data are analyzed by computers in Manhattan and compared with the outcomes of other learners. In the end, a personal learning plan for the next day is created for each student.

      The successes have been impressive. Before Teach to One arrived at Boody in the 2010/11 school year, the performance of the sixth-graders was just below the average of comparable schools in New York City. Two years later, the youngsters’ knowledge level was already 11 percent higher. By now, New Classroom students are learning 50 percent more per year than the American average. Boody hardly enjoys the advantages found at more privileged institutions. The student body is ethnically mixed with many newly arrived in the US. Parents often receive social assistance, 80 percent of the children are entitled to a free school lunch for financial reasons.

      That makes Joel Rose, co-founder and chief executive officer of New Classrooms, all the more proud. “Our team has achieved what nobody has ever achieved before: We provide an individual curriculum for thousands of children – every day anew.”¹² Thanks to its algorithms, New Classrooms reconciles what was previously irreconcilable; it offers personalized education even in large learning groups, eliminating the contradiction between mass application and excellence. Its algorithms do not replace teachers, but support them: Instead of imparting standard knowledge, educators can assist individual students or small learning groups. As impressive as the stories of Felix, Bettermarks and New Classrooms are, the outcomes are merely what happens when a conscientious doctor or teacher has sufficient time to provide individual assistance. A good physician personalizes medical treatment for each patient, good teachers do the same for their learners. But today, this luxury is no longer reserved for the privileged few. Vivienne Ming’s algorithm can offer it to millions of diabetes patients every day, Joel Rose’s software to tens of thousands of students. At the same time, personalizing algorithms provide unscrupulous profiteers with considerable room to maneuver.

       Predatory advertising

      “Find out where their pain is.” That was the message on a poster in which a dentist bends over a patient who is obviously in great pain. However, the picture was not found in a dental clinic, but in training materials for the marketing staff at the ITT Technical Institute, a for-profit educational institution based in Carmel, Indiana, in the United States. The slogan was actually a call to acquire new customers, especially people in crisis. Algorithms helped the salespeople find them.

      Using software, universities such as ITT were able to disseminate their advertising directly to people who searched the Internet for certain words or combinations of words. The sales staff at ITT took the message to look for vulnerabilities seriously. They identified users who, based on their search queries, seemed to be single parents, newly divorced, dissatisfied at work, or lacking self-confidence – at best, all at once.

      In such cases, the institute surmised, people are particularly open to a new start – i.e. a study program that supposedly offers excellent career prospects and the chance to get ahead. This is exactly what the school’s advertisements promised. Those who were desperate enough to click on them and leave their telephone number were immediately called. In some cases, up to 180 times in one month. The goal was to get a signature on a contract for a costly educational program – even if the program was unsuitable, ineffective and of questionable reputation. This aggressive marketing strategy, which focuses not on the quality of the product but on the vulnerability of potential customers, is employed by several for-profit universities in the US. And it seems to pay off. The Apollo Group, for example, whose institutions such as the University of Phoenix have had up to 600,000 students enrolled, spent about $2,200 per student on marketing and only about $900 on teaching each year. In comparison, Portland Community College, a public institution, invests nearly $6,000 in teaching and only $185 in advertising.

      Cathy O’Neil describes all of this in her book Weapons of Math Destruction.¹³ O’Neil’s life is like the story of Saul becoming Paul. The mathematician, who received her doctorate at Harvard, first embarked on an academic career before moving to the financial sector. Her personal turning point was working for a hedge fund. There she experienced mathematical models as both initiators and accelerators of the crises on the housing and financial markets. To her, big data seemed merely to be a springboard to a ruthless and previously unknown form of profit maximization. СКАЧАТЬ