A new Interview around the computer: Thiebaut Weber, trade unionist CFDT and former student activist. He was the Confederal Secretary of the European Trade Union Confederation between 2015 and May 2019. He is interested in digitization, new forms of work and online platforms.
Binary: could you tell us about your career?
I was a student unionist, president of the FAGE. I did not finish my studies: the CFDT offered me to take union responsibilities. I have just finished a four-year term at the European Trade Union Confederation, in Brussels. This confederation brings together the main European trade unions, to act on European employment and social policies, to represent workers and to reflect on issues related to the future of work with the European Trade Union Institute (ETUI). We are obviously interested in the transformations of work by computers, and in the relations between humans and machines. I was also part of a group of European experts on AI where I was the only trade unionist. If I did not personally follow a science course in computer science, I trained myself on the job, especially through contact with friends and experts. I will now join the Interministerial Delegation on Prevention and the Fight against Poverty, where I intend to focus particularly on youth poverty.
The digital world is transforming society. What are the real issues?
Perhaps the most important issue is citizen training and popular education to become a citizen and employee in a digital world. We are at a key moment when we can still define what framework we want for human-machine collaboration. Currently, debates often focus on topics that are science fiction and not enough around the real issues for example: how much do we want to use new technologies to improve our framework or how can we protect data from each beyond the RGPD. These are real societal issues and we need to develop union training to be ready for a dialogue that is not only social but also technological.
For example, artificial intelligence (AI) systems used in workplaces must be irreproachable in terms of protecting human dignity and safety. To do this, it is necessary to establish a dialogue from the design of these systems to exchange on all facets of their deployments in workplaces. Take a building in a chemical plant, connect a system with cameras to permanently locate the hazardous materials; this should not be overly used to monitor workers. We want an ethical AI, a reliable, explainable, and trustworthy AI, acceptable to all. Unions must participate since the design of such systems. Trade unionists must, therefore, be trained to participate in discussions about the uses of such systems.
To take another example, the data produced by the workers, what is done about it? Once they have been used for a particular service, what happens to them? The employer has, to a certain extent, a right to use these data for the supervision of the work done, but this use must be proportionate, remain within certain limits. We need to be able to control what they are used for. It is at this price that the use of the data becomes acceptable by the workers, that confidence is established. In this context, employee training is essential.
Obviously and this is not enough of a subject in the business. We need computer education for our citizens, continuous training in the company. For example, in vocational training, we need training on the use of data, the stakes of human-machine collaboration. We must, in dialogue with scientists, offer new training for the use of these technologies that ensures the competitiveness of our companies, of course, but also in the service of improving working conditions.
This is a good subject that the social partners should take ownership of in the years to come. Social dialogue in France is a little stalled. Digital training could be a subject to boost it by preparing France for the challenges of tomorrow. It seems abstract, but it is very concrete: to go to successful companies and where it will work well, it requires training in new technologies in companies.
What about the disappearance of trades, and the appearance of new professions?
We have already gone through three industrial revolutions. This continues with computer science and artificial intelligence. There is a scientific consensus around the figure of 10% of jobs that should disappear, but no visibility on the new jobs that could be created. For me, the main issue is the evolution of trades. First, in their content: simple tasks of interpretation and decision support will be taken by machines tomorrow. Then everything depends on choices at the company level. For example, let’s take the insurance sector, claims, currently handled by humans. If tomorrow they are treated by OCR, the company can take the opportunity to make a massive layoff plan. It may also choose to implement a training plan for its employees replaced by software to train them in insurance consulting. It is, therefore, necessary conscious actors, enlightened technological stakes to make the right choice. A bad choice can be catastrophic for the company. Unthinking automation in addition to being catastrophic for the jobs can also result, one realizes it more and more, in degradations of the quality of the services or the products.
Unions are there to support the transformations of work. Are they ready for that?
Trade unionism has always been much more comfortable with large collectives than with smaller entities. Today we are witnessing a form of balkanization of work, explosion of units. There are more and more small businesses, independent workers, and that does not promote union work. How to offer the union tool to workers who are isolated and therefore more vulnerable to the risks of exploitation and precariousness? Unions are not organized for these new challenges. In traditional enterprises with appropriate trade union structures, trade unionists can learn about these new technologies.
They need to invest these new subjects in large structures, large administrations, health, and hospital, for example, to start there.
At the other end of the spectrum, let’s take the example of VTC drivers like Uber. Unions have established themselves. A driver, the union activist, has a lot of difficulty in entering into a classic social dialogue with the platform. For example, he needs to know how the algorithm attributes races or not. But there is no explanation, transparency. Union drivers report that they have experienced a slowdown in the races they are allotted. How can a driver prove that he has been discriminated against by the race allocation algorithm? If a judge is interested, he will have to ask experts who will have access to the code, data, test the algorithm … There is no legislation that protects these drivers, nor even a way to check what the platforms do.