In the automation and digitalisation we experienced hitherto, people were given a machine—such as a laptop with the usual office software, a 3D printer or a computer-numeric-controlled milling machine—which they could use to perform their job. Knowledge and communication became more mobile. At the same time, the new machines made possible customised production.Artificial intelligence (AI) systems enable machines to work with people. As they are being introduced into the workplace, new kinds of co-operation are already being defined. And there is little doubt AI systems will play a far greater role in people’s lives. In the future, machines could predict errors or disruptions in work processes (for example, in the context of predictive maintenance) or conduct the beginning of a phone call with a customer.These changes require adaptations. But who will have to adapt? Who will determine which adaptions are made and what form should they take?
Expectations placed on workers
We often hear that workers should receive continued training, to remain ‘employable’. And who has anything against training? Yet the debate isn’t progressing beyond general demands. How individual professions and sectors are affected, as well as the corresponding satisfaction of the demand for training, have not been adequately addressed or implemented. Instead, expectations are placed on workers to behave in an economically rational way—and to kindly get some training.
Participation in continued professional training programmes has in fact risen since 2010 and while now stagnating is at around 50 per cent. On-the-job training enjoys particularly high acceptance. But it is also clear that not all workers are being reached. Individuals with less formal education, those in smaller firms, those who are older or who work part-time participate less in continued training.
One study actually
showed that those workers who can be easily replaced are the least likely to participate in training. In a way, we are perpetuating the inequalities of our school system. We are in danger of of ending up with an even more divided labour market, with well-paid specialists on the one side and a new precariat which performs ancillary tasks—before and after the algorithm—on the other.
To ensure that current changes lead to more rather than less social cohesion, we have to create contextual conditions that make workers feel protected from unfulfillable demands. After a 40-hour working week, most of us do not have the time alongside family and care work to take part in a training programme.
This will become even less likely if labour gives into demands from the employer lobby for more flexible working times and a softening of free time. We must create more time and space to empower and protect workers who—for whatever reason—don’t want this. Here legislation, such as in Germany the Qualification Opportunity Act and the Work-for-Tomorrow Act, as well as in-company guidance on continued training, should play an important role.
With these legislative initiatives, Germany’s labour minister, Hubertus Heil, has already proposed or begun to enact improvements. Under discussion are the many aspects of how to finance continued training programmes. In the Work-for-Tomorrow Act, continued training is made extremely attractive for companies affected by structural change.
Workers don’t just need subsidies—they need time and guidance. And the Federal Labour Office has been offering continued training advice since the beginning of 2019. A right to continued training should guarantee that employers provide enough time for it. Under such conditions, programmes can be created that
Works councils are key actors making sure that such opportunities are actually used. They are not ‘inhibitors’, trying to prevent the introduction of AI systems. Rather, they ensure that processes are implemented well and that tasks are redistributed. They do the preliminary work that will lead to greater acceptance of AI systems within companies, while at the same time providing better working conditions for the workforce.
For this to happen, we need works councils that are knowledgeable about the material. At the same time, works-council members, especially those who are not exempted from their regular tasks to take care of their works-council duties, have enough on their plate. We cannot simply unload additional tasks on to them. So they must be able to bring external expertise, on AI, data privacy and additional aspects of digitalisation, into the workplace.
Some employers are already engaged in active union-busting. In future, it will be even easier to hinder the activities of works councils if workers see and talk to each other less and less, because they will ever more frequently be working at different times and in different locations. A sense of belonging and exchange with works councils can be weakened this way.
As a reaction, work councils must also become more digital. As companies are redefined, we also need new forms of organising works councils—for example, by organising the next election via a Messenger group. We want to address such issues and others in an amendment to the Works Council Constitution Act.
Shaping the transformation
But which abilities and skills, which kinds of knowledge will broad segments of the population—such as, for example, a 56-year-old steel worker, a 32-year-old father working part-time or a 61-year-old tax adviser on a temporary contract—need in the future, to help shape the transformation and get closer to the goal of good work?
Already today, human-machine co-operation places new demands on workers. How will we deal with our algorithmic or robotic colleagues?
Complete trust in an AI system or dealing blindly with decision processes—similar to the annoying clicking away of cookie settings on websites and the acceptance of unread terms and conditions—should not be what we see in workplaces. For this reason, the ability to
think critically and to question results is all the more important.
Such ability must be founded on basic knowledge of how AI systems work. This does not mean everyone needs to be able to write code. But we should learn that AI systems and their decision-making processes have strengths and weaknesses. A successful instance of the imparting of knowledge is the ‘
Elements of AI’ online course, now available in a number of EU languages, including German.
Technical knowledge also remains relevant because it empowers workers to provide critical feedback on the implementation and continued development of AI systems—because not each and every technical possibility can be practically and reliably implemented. Technical knowledge keeps workers on an equal footing with the algorithms.
Besides good formal training, tomorrow’s workplace will require from workers more communication and teamwork, because the tasks under discussion will only be able to be mastered by a group. Here, workers’ ability to give and receive constructive criticism will be key to maintaining their ability to keep learning.
What are called ‘soft skills’ today could soon prove to be ‘hard skills’, essential for companies’ success. Such skills are formed and fine-tuned through their daily application. Both employers and employees face the challenge of organising work in the future in a way that will foster these indispensable skills.
Through works councils and staff committees, workers must be integrated into the planning of continued training. This includes the development of training programmes and the form they will take, and the preparation of plans for a qualified workforce.
It is the duty of social democracy to make sure that human beings don’t get the short end of the stick in the human-machine partnership. At the end of the day, machines should enable more autonomy and bring us closer to the goal of good work. We do not need another relationship of dependency in which the machine is constantly telling us where to go and which movements we should make.
We’ll do that ourselves. To be able to codetermine the shape of the workplace of tomorrow, we need new contextual conditions, of which education and continued training are part—but only part.