Who Trains the Data for European Artificial Intelligence? - ENSAE Paris Accéder directement au contenu
Rapport Année : 2024

Who Trains the Data for European Artificial Intelligence?

Résumé

Executive Summary Microworkers: The Forgotten Platform Workers On February 22, 2024, a panel of microtask data workers (”microworkers” for short), assembled by the EnCOre initiative, addressed grievances at the Fourth Transnational Forum on Alternatives to Uberization, held at the European Parliament. This event convened platform workers’ unions, advocacy organizations, and policymakers to advocate for the adoption of the Platform Workers Directive. Attendees at the Forum expressed unequivocal solidarity with the microworkers’ representatives, leading to the integration of their specific demands into subsequent public outreach efforts and documentation. Microwork refers to low-paid occupations, such as recording one’s voice, translating short bits of text, and classifying contents in an image or webpage. These tasks are essential functions in the development of machine learning and AI because they prepare and correct algorithmic solutions. The Covid-19 pandemic, marked by lockdowns and job losses, along with technical phenomena like the advent of generative AI, has led to a significant rise in microwork across Europe. Microwork in Europe: What We Know So Far Based on our previous survey of more than 4,000 small and medium-sized businesses in two European countries, up to three out of four companies are aware of these platforms. Microwork platforms are increasingly utilized by companies in Europe to recruit, without committing to longterm contracts, a contingent workforce specializing in data-related tasks. Estimates of the global population of microworkers vary, ranging from overinflated platform reports claiming tens of millions of workers each, to more conservative but still alarming estimates by the World Bank, which suggest that this phenomenon accounts for between 4.4% and 12.5% of the global labor force. Estimating the growth of microwork in Europe is challenging due to high turnover, evolving corporate practices, and trade secrets. Recent studies show a general increase in this type of platform work. Although larger EU countries have more microworkers in absolute terms, smaller countries exhibit a higher density. Additionally, significant offshoring to regions like Africa, South-East Asia, and South America may have shifted demand dynamics for European microworkers. Recent data shows that while the majority of European businesses are aware of microwork platforms and data work deep labor networks, only a minority, mainly small businesses, admit to using them. Small businesses engage in microwork indirectly through purchased digital or AI services, with direct microwork clients being primarily technology suppliers. The EnCOre Initiative The EnCOre project, has been engaging platform users, workers of business process outsourcing (BPOs) firms, and technology companies through interviews, focus groups, and surveys in nine European countries: Austria, Belgium, France, Germany, Ireland, Italy, Luxembourg, Portugal, and Spain. Between March 2023 an April 2024, we have conducted exploratory interviews (N=7), focus groups (N=18), and fielded online questionnaires (N=577). Results Our survey reveals distinct demographic trends among microworkers, showing a significant overrepresentation of younger people with high education levels. The median age of participants is 27, with an average age of 30.87 years, indicating a younger demographic compared to previous international surveys. The significance of this trend can be seen in the fact that 37% of respondents possess a bachelor’s degree and 27% hold a master’s degree or equivalent. A high level of education among young microworkers corresponds with the recent increase in the percentage of 25- 34 year-olds obtaining higher education degrees throughout the OECD. Among our respondents, women account for 41% of microworkers. In Europe, there is a relatively smaller gender gap 3 when it comes to the participation of women in microwork. Microwork and Unemployment Portuguese, Italian, and Spanish microworkers make up the largest percentages of participants among the nine targeted countries, with respectively 33%, 20%, and 14%. Notably, all of these countries have higher unemployment rates than the average for the Euro area. While 34% of respondents hold a primary, permanent salaried job, 28% are self-employed or in temporary employment. About 25% of those who engage in microwork do so because they lack other paid work. Of these, nearly 15% are unemployed and actively looking for work—double the 6.4% unemployment rate in the Euro area. It is estimated that less than 20% of microworkers depend on platforms as their primary source of income. However, this small but active group utilizes an average of 2.46 platforms, demonstrating the critical role these platforms play in their financial ecosystems. Despite microwork being a crucial revenue stream, also for those who use it as a secondary source, 75% earn less than 50 euros per month, with median earnings ranging from 20 to 25 euros. Moreover, interviews and focus groups revealed widespread dissatisfaction among workers regarding wage theft, with payments not only being irregular and unsatisfactory but also being withheld arbitrarily. Microwork and Migrations The results of our investigation demonstrate the significant participation of migrants in the European microwork market. 21% of the respondents were born outside of their country of residence, and 18% hold a nationality other than that of their country of residence. There is potential evidence to indicate that migrants are engaging in online labor platforms at higher rates than previously recognized. This is a departure from earlier findings that predominantly described migrants in Europe as mainly involved in locationbased platforms such as transport and deliveries rather than online freelancing and microwork. 59% of the respondents rely on a single platform for their income, suggesting a form of economic dependency, while 41% spread their efforts across an average of 3.42 platforms, indicating a strategy to diversify their income streams. It is noteworthy that almost 90% of these workers also engage in other online earning activities, internet sales, cashback operations, and participation in panels, which do not require specialized skills, and can be conducted on a temporary basis. The above scenario contrasts with the concern expressed during focus groups that many migrants have high educational levels but are unable to find jobs or earn good wages in their new countries. Many are confined to low-paying microwork despite their qualifications. The disconnect raises important questions about the integration and utilization of competencies among migrant workers in the digital economy in Europe, where they still seem unable to find highvalue online freelancing jobs requiring computer programming, design, and project management expertise. Moreover, fewer microworkers are involved in paid content creation and influence, which require significant time investments or specialized training, emphasizing the difficulty they face in leveraging their educational background into more lucrative online employment opportunities. The Human Cost of AI The EnCOre study reveals that despite the trend towards market polarization where highvalue AI startups are located in the Global North and data workers in the South, there is strong development of a microwork ecosystem in Europe. In accordance with our framework, which distinguishes AI preparation, verification, and impersonation microtasks, we provide evidence that European microworkers are heavily engaged in AI-related tasks. During AI preparation, workers generate or annotate data for machine learning applications, such as recording voice prompts or labeling images. During AI verification, workers are required to assess the accuracy of AI-generated outputs, including audios and transcriptions. AI impersonation tasks, where human workers simulate artificial processes, are rare and typically performed internally. In these companies, workers are bound by non-disclosure agreements (NDAs), a practice that is uncommon among respondents to our survey recruited through online platforms, but ubiquitous among workers in BPO companies. The use of these agreements restricts the freedom of expression of workers, a situation made worse by the continuous monitoring of their online activities for the purpose of safeguarding proprietary information and assessing their performance. The psychological and emotional effects of these microtasks are profound on workers. They are often referred to as merely ”data tasks” but entail significant emotional engagement and pose significant socio-psychological risks. This phenomenon is particularly evident among microworkers involved in content moderation, who frequently report adverse effects on their mental health. This type of work has the potential to result in severe consequences due to its intense 4 nature. Among them are incidents in which workers have fainted, suffered from burnout, experienced psychotic episodes, and, tragically, in at least one instance, committed suicide. Moreover, the isolation experienced by these workers is further exacerbated because of remote work or, in the case of shared workspaces, a general distrust among colleagues and corporate management. The combination of high-stress tasks and social isolation presents a challenging environment for microworkers, highlighting the need for regulatory and support mechanisms in order to mitigate these impacts. Policy Recommendations As policy-oriented initiatives predominantly focus on location-based platforms, microworkers are not adequately represented. Owing to the dispersed nature of their activity, they face psychological rather than physical challenges owing to the social isolation arising from the dispersed nature of their activity, and in the case of content moderators, from exposure to all sorts of online toxicity. They are also more vulnerable to disclosures of their personal data and privacy violations. Low remunerations, limited direct communication with clients and platforms, and difficulty to challenge payment withholdings are common complaints, but the pervasive misconception that microworkers choose this type of work voluntarily discourages efforts to improve working conditions. This report highlights that microworkers are indeed platform workers, and stresses that their distinct needs must be acknowledged in discussions and legislation pertaining to platform work.
Fichier principal
Vignette du fichier
Report_EnCOre-final.pdf (7.47 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
licence

Dates et versions

hal-04662589 , version 1 (26-07-2024)

Licence

Identifiants

  • HAL Id : hal-04662589 , version 1

Citer

Milagros Miceli, Paola Tubaro, Antonio A. Casilli, Thomas Le Bonniec, Camilla Salim Wagner, et al.. Who Trains the Data for European Artificial Intelligence?: Report of the European Microworkers Communication and Outreach Initiative (EnCOre, 2023-2024). European Parliament; The Left. 2024, pp.1-40. ⟨hal-04662589⟩
84 Consultations
26 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More