The AI at the Siemens factory will distribute tasks to robots and people programming their joint work
If You once came up with the idea that the project manager is a practically useless post, and you can automate it, then you were right. Siemens did this.
Many modern factories use industrial robots. Often there are more robots on the conveyor belt than people. The main problem of this automation is that the robots need to be reprogrammed for each job. It is still necessary to repair the faulty mechanisms, and only a person can do it, even with a risk to life. In any case, people and robots are destined to work together. So, Siemens is experiencing a new configuration of such cooperation: at its "factory of the future" people and robots work together under the control of a computer program (you can call it the weak form of Artificial Intelligence).
Siemens AG is a German concern that engages in mechanical engineering and electrical engineering. The company spends 366,000 employees, dozens of factories in different countries, and has an annual turnover of € 80 billion. The company spends almost € 4 billion a year on research and development, which is comparable to the Russian federal budget for science (336 billion rubles in 2017). Given the scale of Siemens' activities and the number of factories, the industrial giant is extremely interested in maximum automation of production.
So, industrial robots are working. It's time for Automation 2.0, where robots and people are effectively controlled by Artificial Intelligence, writes Fast Company . AI programs robots and assigns tasks to people.
"Instead of programming each robot, we say that this machine can do such a job, but this machine has the following options," explains Florian Michahelles, who Leads the Siemens Web of Things research team in Berkeley (USA). His group developed a program called a "reasoner". This is a kind of expert system. It independently determines what steps must be taken to produce a particular product, then it distributes tasks among robots in accordance with the capabilities of each of them.
According to the calculations of the researchers, the automated factory UberManufacturing will work "on demand", like the Uber taxi service . Unlike modern mass production plants, this factory can accept small unit orders – and reconfigure robots on the fly.
Slides from the presentation of Florian Mikaels, where The theme of an automated factory with a "reasoner"
Siemens is one of the first companies that integrates modern robots and programs of Artificial Intelligence. Some researchers believe that the future behind such integration of robotics, AI and machine learning. Analysts at McKinsey Global Institute believe that the increase in productivity from automation in the coming decades will significantly exceed the average annual increase in productivity from the steam engine in 1850-1910 (0.3%), from early robotics in 1993-2007 (0.4%) and information Technologies in 1995-2005 (0.6%). In 2015-2065, automation will increase productivity by 0.8-1.4% per year. For a developed Western economy, this is a very large growth.
Preliminary analysis shows that only 5% of staff will lose their jobs due to the full automation of their activities (probably project managers can be among them). But automation will partially affect virtually every profession: from miners and landscape designers to bankers, welders and executive directors.
The most vulnerable to the automation of the profession is related to physical labor, which is performed in a well-structured and predictable environment. The potential of automation by profession is well represented in the interactive tool of McKinsey for 750 professions. There are shown professions with an automation potential of 1% (financial advisors) to 100% (operators of loading machines).
The Siemens research team has already conducted an experiment and proved that an automated factory controlled by AI can work in Small scale: the test system used several types of robots and assembled five types of furniture with four configurations of the legs, six color variants and three types of floor protectors (total 360 combinations of the possible product). Thus, the automated factory will be able to execute personalized orders in small batches.
Although the system provides for the integration of human labor and robots, experts understand that the proportion of people in factories will gradually decline. In the future, robots can perfectly cope with all tasks without us, including repairing damage and reproducing themselves. So under AI management such a factory can work continuously.