Beyond the Black Box: Why Fairness, Not Transparency, Is the Key to Algorithmic Management That Works for Platform Workers

Finland must transpose EU Directive 2024/2831 on platform work by December 2026. The directive operates on an intuitive logic: make the algorithm visible, and workers will be better off. However, transparency and fairness are not the same thing; knowing how an algorithm works does not guarantee it treats you fairly.

The limits of disclosure

Directive (EU) 2024/2831 requires platforms to inform workers about automated systems and provide human review for significant decisions. While this is a necessary step, the operative articles lean heavily on disclosure under the assumption that understanding a system makes it feel more equitable.

Recent studies challenge this assumption. Research shows that providing too much information can actually increase worker resistance and make the system feel less fair (Hu et al. 2024). Furthermore, workers tend to judge fairness almost entirely by whether their pay feels proportionate to their physical effort, meaning an understanding of the underlying process barely registers: a phenomenon labeled the transparency fallacy (Mirbabaie et al. 2025).

Organizational justice theory explains why. Colquitt (2001) identifies four dimensions that workers evaluate: procedural, distributive, informational, and interpersonal justice. Informational justice is only one component. Distributive justice, the sense that outcomes match effort, is the dominant driver of satisfaction. This is reinforced by the Job Demands-Resources model (Bakker & Demerouti 2007), which shows that fairness sustains motivation in ways transparency alone cannot. Ultimately, understanding an unfair system does not make it fair.

Design choices and technical ceilings

Modern dispatch algorithms may be genuinely impossible to explain. Because machine learning systems evolve on real-time data, their behavior is difficult even for their designers to interpret (Möhlmann et al. 2021). Consequently, transparency mandates may hit a hard technical ceiling before workers see any real benefit.

There is also a fundamental question transparency cannot answer: what is the algorithm optimizing for? For example, Wolt’s documentation notes that its system routes orders by proximity and efficiency. It is explicitly not designed to distribute income equitably among active couriers.

Researchers at IIT Delhi found that couriers working identical hours earned significantly different amounts due to dispatch logic rather than effort. Their alternative, equity-focused model, FairFoody, achieved up to ten times more equitable income distribution with minimal impact on delivery times (Gupta et al. 2022). This proves that ignoring earnings equality is a deliberate design choice, not an engineering constraint.

The reality on the ground

From the courier’s perspective, official transparency reports do little to solve daily operational uncertainties. During over-saturated shifts, the distribution of work feels entirely arbitrary. Reading published documentation does not explain why one courier’s order counter sits at zero while nearby courier receive steady deliveries. As one courier noted in a 2022 AlgorithmWatch investigation, the algorithmic process still feels like a secret the platform is keeping (Kayser-Bril 2022). Couriers do not just want data clarity; they want predictable income and a way to challenge automated decisions.

This preference is being reflected in reality. A bachelor’s thesis study at LAB University of Applied Sciences surveyed 40 Wolt couriers in Finland in April 2026. The study revealed that perceived fairness was the strongest predictor of job satisfaction (r = .802), remaining dominant even when transparency and monitoring were statistically controlled for (Mujahid & Farhan 2026). Perceived transparency showed no independent predictive power once fairness was accounted for, indicating that transparency only matters as a vehicle to achieve fair treatment.

A path forward for Finland

Finland’s upcoming transposition is an opportunity to go beyond the directive’s minimum requirements. The Finnish Supreme Administrative Court’s ruling KHO 2025:41, which classified Wolt couriers as employees, already signals that algorithmic supervision constitutes an employment relationship in substance.

To build on this mandate, the Finnish framework should introduce:

• Genuine contestation rights: A real mechanism to challenge automated decisions and receive human intervention from someone with the authority to correct errors.
• Third-party audits: Regular assessments of distributional outcomes to ensure earnings are equitable across the worker population.

Generative AI was used for language editing, literature search support, and research synthesis during the preparation of this article. All AI-generated content was reviewed, revised, and verified against primary sources by the author.

References

Bakker, A.B. & Demerouti, E. 2007. The job demands-resources model: State of the art. Journal of Managerial Psychology, 22(3), 309–328.

Colquitt, J.A. 2001. On the dimensionality of organizational justice: A construct validation of a measure. Journal of Applied Psychology, 86(3), 386–400.

Directive (EU) 2024/2831 of the European Parliament and of the Council of 23 October 2024 on improving working conditions in platform work. Official Journal of the European Union, L, 1 December 2024.

Gupta, A., Yadav, R., Nair, A., Chakraborty, A., Ranu, S. & Bagchi, A. 2022. FairFoody: Bringing in Fairness in Food Delivery. arXiv:2203.08849. Cited 27 May 2026. Available at https://arxiv.org/abs/2203.08849

Hu, P., Zeng, Y., Wang, D. & Teng, H. 2024. Too much light blinds: The transparency-resistance paradox in algorithmic management. Computers in Human Behavior, 161, article 108403.

Kayser-Bril, N. 2022. Wolt: Couriers’ feelings don’t always match the transparency report. AlgorithmWatch, December. Cited 27 May 2026. Available at https://algorithmwatch.org/en/wolts-algorithmic-transparency-report/

KHO 2025:41. Korkein hallinto-oikeus 22.5.2025, taltionumero 1117/2025. Cited 27 May 2026. Available at https://www.kho.fi

Mirbabaie, M., Langer, M., Rieskamp, J. & Hofeditz, L. 2025. Transparency fallacy: Perceived fairness in algorithmic management. Business and Information Systems Engineering. Cited 27 May 2026. Available at https://doi.org/10.1007/s12599-025-00963-1

Mujahid, M.N. & Farhan, M.S. 2026. Algorithmic Management and Courier Platform Work: A Case Study of Wolt in Finland. Bachelor’s thesis. LAB University of Applied Sciences. Cited 27 May 2026. Available at https://urn.fi/URN:NBN:fi:amk-2026060321046

Möhlmann, M., Zalmanson, L., Henfridsson, O. & Gregory, R.W. 2021. Algorithmic management of work on online labor platforms: When matching meets control. MIS Quarterly, 45(4), 1999–2022.

wal_172619. 2023. Sade, kuljetuspalvelu, kuriiri. Pixabay. Cited 5 Jun 2026. Available at https://pixabay.com/fi/photos/sade-kuljetuspalvelu-kuriiri-7803539/