The AI ​​shows its effect immediately

People want AI - and are also afraid of it

  • One in two expects AI to change society noticeably
  • 9 out of 10 require secure AI applications, but only half as many want bans

Berlin, September 28, 2020 - More and more people are discovering the opportunities of artificial intelligence and would like it to be used in almost all areas of life. A majority anticipate that AI will change society noticeably in the next five years and are calling for the technology to be controlled more closely. That is the result of a telephone survey of 1,004 people aged 16 and over on behalf of the Bitkom digital association. Accordingly, only 5 percent say that they have never heard or read anything about AI. Two years ago it was 12 percent, three years ago even 22 percent. More than half (52 percent) are convinced that they can at least explain what AI means (2018: 40 percent). At the same time, more than two thirds (68 percent) now say that they see AI primarily as an opportunity. Three years ago it was only a minority at 48 percent; in 2018 the proportion was 62 percent. "Those who are better informed also see the opportunities of artificial intelligence," said Bitkom President Achim Berg today at the presentation of the study.

Three quarters would like a German leadership role in AI

The increasing knowledge about AI is also likely to be related to the fact that a large majority already use AI applications in everyday life. At the top are text suggestions for writing messages (68 percent), route suggestions for navigation (62 percent) and voice assistants on smartphones (60 percent). But also recommended titles for streaming (44 percent), automatic translations (42 percent), driver assistance systems in the car (39 percent) or purchase recommendations in online shops (34 percent) are already being used deliberately. One in five (20 percent) uses face recognition to unlock their smartphone, and one in eight (12 percent) uses face recognition for photos to make it easier to find people.

Citizens see AI as being of paramount importance for the economy and prosperity. Two thirds (66 percent) assume that AI will strengthen the competitiveness of the German economy. And three quarters (75 percent) demand that Germany take a global leadership role in the development and marketing of AI applications. At the same time, a large majority (85 percent) want secure AI and demand that AI systems in Germany be checked particularly thoroughly and only be used in devices once they have been approved. Only a minority of 44 percent would like to ban certain AI applications in Germany. “People are realizing the crucial role AI will play in the future. Instead of bans, they want control and security, ”says Berg. “AI applications are already subject to existing regulations such as anti-discrimination and data protection regulations. Now it is necessary to consistently implement specifications for particularly safety-critical products and, if necessary, adapt them to AI-specific risks. We don't need new laws for AI. "

More AI is particularly desired in nursing, administration and medicine

A majority would like the use of artificial intelligence in almost all areas of life. At the top is nursing (75 percent), for example to monitor the health of older people. This is followed by offices and authorities (73 percent), medicine (67 percent), security (66 percent) and sport (61 percent), for example in referee decisions. A majority would also like the use of AI in transport (58 percent), in education (55 percent), in HR departments or in customer service (54 percent each) and in traffic (52 percent), for example with autonomous vehicles. On the other hand, AI does not find a majority for political decisions (50 percent), in the military (48 percent) or in looking after small children (38 percent). "Artificial intelligence is particularly desired where the personal benefit is tangible - and where there is possibly the greatest need for improvement: in nursing and medicine as well as in administration," says Berg. "But it also becomes clear that, according to the citizens, AI has potential in almost every area of ​​life."

In the world of work, fears of control and job loss dominate

People are skeptical about using AI in their everyday work. 44 percent see the main dangers of AI in this area. Around three quarters (73 percent) fear greater control of employees by AI, two thirds (65 percent) fear the loss of jobs. Almost one in two (45 percent) believes that AI helps avoid mistakes in the job. Just as many assume that boring routine activities will be reduced, leaving more room for more interesting tasks.

When it comes to the use of AI in personnel selection, the population is divided. One half (49 percent) fear that an AI will reject an applicant for no objective reason, for example because the algorithm discriminates against certain people. On the other hand, at 44 percent, almost as many say that they are more likely to fear that a human HR manager would reject applicants for no objective reason, for example because they have prejudices about age, gender or origin. Berg: “The intensive public discussion about discriminatory algorithms is reflected one-to-one in the survey results. AI is almost impossible to find in HR departments. Less than one percent of companies currently use AI to select applicants. "

AI warning systems and autonomous vehicles will shape our mobility

There are particularly high expectations for the use of AI in traffic. Three quarters (76 percent) assume that AI-supported warning systems in cars will have established themselves in the next ten years, 38 percent even expect a breakthrough within five years. Two thirds (66 percent) expect that within ten years AI will plan optimal routes for us across different modes of transport. There are very different expectations for self-driving vehicles. 60 percent expect self-driving buses to break through on our roads within ten years. But only 44 percent expect self-driving underground or suburban trains, 39 percent long-distance trains and 37 percent self-driving delivery vans for the transport of goods. 30 percent can even imagine autonomous cars on German roads within ten years, 8 percent even expect autonomous cars to be on the road within the next five years. The situation is very different in aviation: just 6 percent believe that autonomous aircraft will transport passengers in the next ten years.

Citizens do not only see advantages in AI-controlled cars. Only one in three (36 percent) says that self-driving cars would prevent serious accidents and many deaths. But significantly more (57 percent) assume that self-driving cars will lead to more serious accidents with many deaths. “In the past, almost every major accident involving a self-driving test car has been extensively reported. Obviously it has gotten stuck in people's minds, ”said Berg. “The additional safety is one of the most important reasons for the development of autonomous vehicles.” As the official accident statistics show, 9 out of 10 accidents with injuries or fatalities in Germany are currently caused by human error, such as excessive speed, insufficient distance or Disregard of the right of way. Berg: "Even if the AI ​​does not immediately control the vehicles completely accident-free, the majority of such serious accidents could be avoided."

AI will advance medical research

In medicine, too, people expect profound changes through artificial intelligence. Six in ten (60 percent) assume that AI will drive advances in medical research in the next ten years. Almost every second person (46 percent) expects AI to provide behavioral advice in order to avoid illness. 42 percent believe that AI will recommend individual therapies for diseases and 40 percent believe that AI will support doctors with routine tasks, such as analyzing X-rays. One in three (33 percent) assumes that AI monitors body data and warns of illnesses or an impending heart attack at an early stage. At the same time, only one in five (20 percent) expects the use of AI in medicine to increase life expectancy in the next ten years, and only 13 percent say that AI will make better diagnoses than doctors. A mere 8 percent expect that AI diagnoses without human intervention will replace a doctor's visit and 6 percent expect diseases that are considered incurable to be cured today. “AI will find its way into medicine just as quickly as it will in mobility. Technologically, there are already applications or at least functioning prototypes for practically all scenarios, ”says Berg. "AI will revolutionize medicine more than the discovery of antibiotics."

Noticeable changes through AI in the next five years

A majority of citizens (53 percent) are certain that artificial intelligence will change society noticeably in the next five years. One in four (24 percent) believe that AI is already doing this today. Two years ago, only 41 percent believed that within five years AI would bring about such changes. "When politicians discuss artificial intelligence, it is often not so much about the technology and its effects, but about more fundamental socio-economic positions and distribution issues," says Susanne Dehmel, member of the Bitkom management and expert member of the inquiry. Artificial Intelligence Commission of the German Bundestag. “The work in the Enquete Commission has shown that this discussion of socio-political issues cannot be resolved with the legal regulation of a technology.” Above all, something that is not yet widely available cannot be regulated in a meaningful way. "Regulation cannot anticipate experience in dealing with a technology," says Dehmel. “We have to think about where the use of AI systems is particularly useful and should be promoted. Risk-oriented ex-ante regulation inherently harbors great risks by closing the enormous potential of innovative AI in Germany. "

Artificial intelligence: data availability must replace data economy as a model

When it comes to the question of whether the use of AI is making progress and whether Germany can take on a global leadership role in artificial intelligence, Bitkom believes that the handling of data is of crucial importance. "As a model, data availability and data sovereignty must replace data economy," said Bitkom President Berg. “We have to create rules in Germany and Europe that facilitate data-based innovations such as AI in the interests of citizens, companies and administrations.” At the same time, demands for independent AI regulation testify to a false image of artificial intelligence. “AI is a technology and, as before, we should also make laws that are as technology-neutral as possible. Regulation should always concern an application and its effects, not the technology as such, ”emphasized Berg. “In cases where AI applications are used in high-risk areas, the existing regulation and the existing institutional structures of the respective sector must be consistently linked. We do not need an algorithm Tüv or, for example, a separate AI financial supervisory authority, but the financial supervisory authority must take AI applications into account. "

At the same time, Bitkom advocates focusing more on application in individual industries in AI research and AI funding. The European Gaia-X project also offers the opportunity to develop the cloud, data economy and AI together. Gaia-X could develop into an open ecosystem that makes it possible, especially in the AI ​​context, to store and exchange data in a trustworthy manner. An AI education offensive is also necessary in order to create a broad social understanding of AI and its effects.

Methodological note: The information is based on a survey that Bitkom Research carried out on behalf of Bitkom. 1,004 people aged 16 and over in Germany were interviewed by telephone. The survey is representative.