Talking about the use of artificial intelligence in labor relations is not commenting on the future, but rather addressing a reality that has been establishing itself in companies for years, including Spanish ones. 8.32% of companies with more than 10 workers were already using this technology in 2021, according to data collected by the INE, while another survey carried out two years earlier by Eurofound raised this percentage above 60% and placed Spain in first position compared to the rest of the EU. However, the information disseminated in this regard within the framework of collective bargaining is scarce.
“The algorithms applied to labor decisions, today, They are a black box and what is worse, it represents an increasingly widespread practice,” points out the Head of Digitalization at Work at UGT, José Varela. This refers to INE data to point out that 40% of these companies use them to automate workflows, while less than 10% say they use them for human resources management. “The problem is that a kind of infallibility is attributed to themwhen they are nothing more than mathematical calculations based on probabilities, sometimes inexplicable, which contain very high margins of error.
Algorithms cannot make a hiring or firing decisiongiven that the European data protection regulation establishes that “every interested party will have the right not to be subject to a decision based solely on automated processing”, that is, that the last word must be had by one person. However, the professor of Labor Law at the University of Valencia and author of ‘Productive and extractive algorithms. How to regulate digitalization to improve employment and encourage innovation’, Adrian Todolimaintains in conversation with this medium that this is not a widely recognized right, which has different implications.
“They are used for recruitment, When you upload a resume to a platform, the first one that reads it is an algorithm that decides whether to move on to the next phase or not and it has been shown that on some occasions they have discriminated against women with the same training or experience,” she explains. “They are prone to generating biases because many times the algorithms are fed by data that is in society and what it does is replicate them, including discriminatory patterns,” says Todoli, who details that The use of these technologies is widespread at work through digital platforms – VTC drivers, food delivery people – but also in logistics centers or even among cleaning staff to measure their productivity.
“If the algorithm calculates that there is an optimal time and the worker takes longer, your device will notify you that you should do it faster. This is a constant surveillance that can end up causing accidents at work if it is not well programmed and does not take into account factors such as rest time or the age of the worker,” explains the Law professor. These records can precipitate the dismissal of part of the staff, but Todoli points out that in some cases it is It is the platform itself that disconnects the worker if they have a bad digital reputation or do not comply with the preset times.
Artificial intelligence in collective bargaining
Law 15/2022 prohibits discrimination arising from the use of AI and automated decision-making mechanisms in the field of public administrations and obliges both these and private companies to promote “ethical, reliable and respectful of human rights” use, indicates the partner in the labor area of Ceca Magán Abogados Àlex Santacana at the information. Employers and unions took charge of this principle in the V Agreement for Employment and Collective Bargaining (AENC) signed in the month of May, which states that companies will inform worker representatives about the procedures used to hiring, evaluation, promotion and dismissal.
The truth is that they are obliged to do so, in accordance with royal legislative decree 2/2015 that expands the Statute of Workers and obligates to communicate to the Company Committee the “parameters, rules and instructions on which the algorithms or artificial intelligence systems that affect decision-making are based.” However, the unions report that they encounter high resistance from companies to transfer this data and point out that It has only been included in a handful of agreements throughout the countrywhich is why both UGT and CCOO are already preparing guides so that their negotiators know what they have the right to demand in terms of artificial intelligence.
“The rider law is a product of social dialogue and is a pioneer in Europe“We also have the European Framework Agreement of the social partners on digitalization in which a complete process is agreed so that the company is not limited to reporting, but also that negotiation with the workers comes into play,” explains Raquel Boto from the Secretariat of Union Action of CCOO. Currently, the European Commission is working on an artificial intelligence directive that will come to light in the coming months, which represents a new advance for the union, although they demand a regulation that specifically deals with the use of this technology in the framework labor.
It has not come to court
“It is a very new issue and at the moment we are not aware that there have been judicial conflicts in this regard,” acknowledges the partner at Ceca Magán Abogados, who assesses that the biases applied by AI can lead to this type of confrontations in the coming years. However, yes There are precedents in the Netherlands. “A Dutch court has ruled that the State has discriminated against its citizens by deciding certain social benefits based on a risk profile that calculated the probability that the application was a fraud. The citizen classified as a potential fraudster was part of a group categorized as problematic, based on their neighborhood of residence and These were monitored automatically“.
In this sense, the collaborator of Sagardoy Abogados and professor of Labor Law, Gemma Fabregat, recognizes some concern among workers about the control that this technology may exert or the discrimination they may suffer. “Personally, what is most worrying is that the algorithm requires you to have the knowledge to know if the solution they are giving you or what they are proposing is true. The problem is that if we reduce the level of knowledge by delegating to these technologies, it could be that what the algorithm indicates becomes the truth because there is no person with the capacity to say that it is wrong.”
The expert defends that history proves that the entry of technology into the workplace cannot be denied, but warns that artificial intelligence affects the unconscious self and therefore, labor law, as his colleague Ignasi Beltran has investigated. “It represents a qualitative leap in history, which leads you to make decisions or carry out movements without you being fully aware of it.” He also warns about it labor partner of AE Abogados, Alejandra Gútiez. “A decision based on an algorithm may exceed the scope of certain fundamental rights or fail to take into account certain subjective variables that escape its programming, which would translate into serious legal consequences for the company that may lead to sanctions or legal actions,” he concludes. .
