It first introduces itself in somewhat choppy but grammatically correct English, saying: “I’m a collaborative robot.” Then it helps a real, live, flesh and blood technician to lift a 30 kg overhead cover. Finally it hands over a spray bottle for cleaning a conveyor belt at a height of more than two meters. ARMAR-6 is not just any robot; the white- and turquoise-colored machine with a human face and five cameras built into its head is equipped with artificial intelligence. Without instructions the robot recognizes when a human needs help, communicates constantly with its human colleague and then independently performs maintenance tasks in industrial plants. “We didn’t program its skills; they were learnt,” says Tamim Asfour, developer of the robot and a professor at the Karlsruhe Institute of Technology (KIT), when presenting the man-machine, which took place at CeBIT in Hannover, Germany, in June 2018. In other words, ARMAR-6 is not limited to specific, preprogrammed activities but can, for example, learn by how to use new tools by observing humans. The professor has no doubt that “robotics and artificial intelligence are key technologies of the twenty-first century and will make a decisive contribution to improving humankind’s quality of life.”
New Megatrend – New Opportunities
One thing is clear: The opportunities, which the use of artificial intelligence opens up for business and society are one of the most-discussed topics currently — far beyond CeBIT. Experts are outdoing each other with optimistic scenarios. A current PricewaterhouseCoopers study, for instance, concludes that Germany’s gross domestic product (GDP) is set to increase by over 11 percent by 2030 due to AI-based innovations alone. That would amount to about EUR 430 billion. According to the management consultants Accenture, around 80 percent of German businesses are planning to automate further business processes on a large scale in the next three years. The Boston Consulting Group also sees significant potentials but notes that there is still a gap between desire and reality. So far, only 28 percent of industrial enterprises around the world are said to have developed a clear AI strategy, while in Germany the figure is only 23 percent.
No matter how big the growth effect may end up being, the bandwidth of AI technologies and methods is certainly wide, ranging from digital assistants and cognitive robots to deep learning systems, or artificial neural networks modeled on the human brain. Their uses are just as varied. Be it driverless transportation systems in logistics, chatbots in customer service or AI-assisted forecasts in finance and accounting. In many areas of use opportunities arise to automate business processes, make them more efficient and to help people make decisions or relieve them of the burden of routine tasks.
The RWE energy group, for example, uses an AI platform to handle customer correspondence. The platform independently recognizes customers’ concerns in emails and forwards the mail to the right contacts. The airport operator Fraport uses adaptive algorithms to analyze passenger flows. That way, queues can be reduced and personnel deployment can be optimized. The agricultural machinery manufacturer John Deere uses smart solutions in production, for example to maintain machines. Their aim is to convert their Mannheim production facility step by step into a smart factory.
Smart, Smarter, AI
John Deere is a typical case. Predictive maintenance of production facilities alone offers significant value-added potential across industries of all kinds. The
Telekom defines policy for the use of AI
basic idea is that large streams of sensor data and other information are analyzed by AI technologies in order to monitor plants proactively. This allows critical states such as overheating to be identified in time. A current PAC study notes that 83 percent of European manufacturing and transportation enterprises are planning to increase their spending on predictive maintenance in the next two years. They aim to minimize repair times and machine outages, to get old equipment going again and to improve customer satisfaction.
Will Humankind Become Superfluous?
In spite of all the optimistic forecasts and promising initial applications in the most diverse of industries, many unsolved issues remain on the way to the world of supersmart machines. Some experts point to imminent job losses. An OECD survey says that computers could take the place of one in five jobs in Germany in the future. Others, in contrast, see opportunities for new professions and qualifications. The market researchers at Gartner, for example, reckon that AI technologies will create more jobs by 2020 than they destroy. Politicians see the challenge and are trying to set guidelines. In mid-July 2018 the German federal government laid the first cornerstones for an “Artificial Intelligence Masterplan,” while in 2017 the Federal Research Minister launched the expert platform “Learning Systems” to promote debate on legal, ethical and technical issues.
According to a current analysis of the manufacturing industry by “VDI/VDE Innovation und Technik”, “the central obstacle to the use of AI is the shortage of specialists and in-house competences.” A PAC study, in contrast, concludes that companies mainly see legal issues and compliance considerations as the most serious impediments. It is not certain, for instance, who will be to blame if an error occurs: man or machine? A further barrier is that in-house processes and corporate culture are in many cases not yet ready for AI. According to PAC, it does not make sense for the companies in question to roll out the new technologies on a large scale from outset. “Instead, it is better to begin with small use cases that achieve positive results fast.”