| Key takeaway: In November 2024, Italy’s Garante fined Foodinho €5 million for illegal algorithmic management affecting 35,000 riders. This decision penalizes unauthorized geolocation tracking and the lack of human oversight in performance ratings. It signals a major regulatory shift, requiring platforms to ensure transparency and human intervention to prevent discriminatory automated decisions under evolving European labor laws. |
Can an algorithm truly replace human judgment when it comes to managing thousands of workers? In November 2024, Italy’s data protection authority, the Garante, announced a €5 million fine against Foodinho, the Glovo group company operating its delivery platform in Italy.
The decision concerned unlawful processing of personal data affecting more than 35,000 riders, including excessive geolocation tracking, insufficient transparency and automated decisions that riders could not effectively challenge.
The case has since developed beyond data protection. In February 2026, Milan prosecutors placed Foodinho under judicial supervision over alleged labour exploitation and worker misclassification, with its CEO under investigation. In May, Glovo agreed to a proposed pay increase of around 40% as part of discussions with prosecutors.
However, by August, prosecutors were still questioning whether the changes adequately compensated riders and addressed their employment status.
What this means that these developments show why platform operators must assess not only how algorithms process personal data, but also how automated management affects pay, working conditions and employment rights.
2026 Judicial Supervision, Alleged Misclassification and Rider Pay
In February 2026, Milan prosecutors placed Foodinho under judicial supervision over allegations of labour exploitation involving approximately 40,000 riders across Italy. The investigation concerned allegations that riders formally classified as self-employed were working under conditions that, according to prosecutors, indicated an employment relationship. Foodinho and its CEO were placed under investigation. These are allegations within an ongoing legal process, not a final determination that every rider must be classified as an employee.
In May 2026, Reuters reported that Glovo had agreed to raise rider compensation by around 40% under a plan negotiated with Milan prosecutors. The proposal included a minimum gross hourly payment of €14, compared with €10 previously envisaged under the relevant sector arrangement, alongside changes to the per-delivery minimum. The plan was intended to address concerns about pay and working conditions while the company remained under judicial supervision.
However, the developments did not immediately resolve the dispute. In August 2026, ANSA reported that prosecutors considered the measures insufficient, arguing that riders were still not being paid for a substantial share of their working time and that their contractual status remained inconsistent with the conditions under which they worked. Glovo disputed this assessment and maintained that it had implemented the agreed changes.
This later phase of the case illustrates why algorithmic management cannot be assessed solely through the lens of privacy. The systems used to assign deliveries, calculate pay, measure performance and determine access to work may also be relevant when authorities examine direction, control and employment status.
For platform operators, the practical lesson is to review compensation formulas, the treatment of waiting and availability time, performance metrics, account restrictions and contractual classifications together. Raising headline rates or improving privacy notices may not resolve wider compliance concerns if the underlying working arrangements remain disputed.
The Garante’s €5 Million Penalty Against Foodinho
On 22 November 2024, the Garante announced a €5 million fine against Foodinho S.r.l., a company within the Glovo group. The underlying decision was adopted on 13 November 2024 and followed inspections into the company’s digital platform and its processing of rider data.
The investigation began ex officio, following press reports about the deactivation of a rider’s account after he died in a road accident while making a delivery in 2022. A report from a group of IT experts also contributed to the investigation. The account incident prompted scrutiny of Foodinho’s wider data-processing practices, including its automated systems, rider monitoring and procedures for suspending or deactivating accounts.
The Garante identified multiple GDPR violations affecting more than 35,000 riders. It ordered Foodinho to change its practices, improve the information provided to riders and ensure that automated decisions could be reviewed and challenged. The authority also prohibited further processing of riders’ biometric data for identity verification through facial recognition.
The case demonstrates how privacy enforcement can expose wider weaknesses in workforce management. Companies operating digital labour platforms should review their systems for lawful data collection, meaningful human oversight, transparent performance assessments and fair procedures for decisions that affect workers’ access to assignments or their accounts.
Violations of the Italian Garante Privacy Standards
In November 2024, the Italian Garante imposed a €5 million fine on Foodinho S.r.l., a key entity within the Glovo group. This enforcement action followed rigorous inspections into their digital infrastructure.
The investigation revealed systemic failures in how the company processed personal data. These illegal practices impacted over 35,000 delivery riders. The scale of the breach suggests a fundamental disregard for privacy standards.
Such failures are closely linked to broader HR compliance issues and challenges. Regulators are now treating these algorithmic errors with increasing severity and scrutiny.
This fine marks a major legal milestone. It serves as a stern warning to other giants operating within the platform economy.
- Total fine amount: €5 million
- Scope: Over 35,000 riders affected
- Parent company: Glovo (Delivery Hero group
- Primary authority: Italian Garante per la protezione dei dati personali
Beyond general privacy standards, the specific handling of location data revealed even deeper issues with worker surveillance.
Unlawful Geolocation Tracking and Facial-Recognition Restrictions
The Garante found that Foodinho’s platform transmitted riders’ geolocation data even when they were not working and the app was running in the background. Until August 2023, location data could also be transmitted when the app was inactive. The company also shared riders’ personal data, including location information, with third parties without providing adequate information to the affected individuals.
These practices raised concerns about necessity, proportionality, transparency and compliance with the GDPR. The Garante required Foodinho to display an indicator when GPS tracking was active and to disable tracking when the app was running in the background. The authority also required the company to review its data-protection information and retention periods.
The decision addressed biometric data as well. Foodinho had used facial recognition to verify riders’ identities, comparing a selfie with an identity document. Although the company had stopped using the procedure in July 2022, the Garante expressly prohibited further processing of riders’ biometric data for identity verification through facial recognition.
The authority also found that Foodinho had failed to meet relevant safeguards under Italy’s Workers’ Statute concerning remote monitoring. Employers and platforms should therefore assess both data protection and employment-law requirements before deploying location tracking, biometric verification or other monitoring tools. A technically convenient system does not automatically provide a lawful or proportionate basis for monitoring workers.
How Automated Rider Management Affected Worker Rights
While the fines focus on privacy, the real operational damage lies in how these algorithms fundamentally altered the employer-employee relationship and HR oversight.
Absence of Human Intervention in Rider Ratings
The Garante decision highlighted a total absence of human oversight. The algorithm alone evaluated and rated worker performance. This created a “black box” for the riders.
We see a clear failure to allow riders to contest results. Workers could not express their views on automated scores. This lack of dialogue is a major HR failure.
This situation directly impacts how to handle poor performance compliantly. Human review remains mandatory for any disciplinary actions.
- Absence of human review in scoring
- Lack of feedback loops
- Inability to challenge automated bias
- Rigid mathematical evaluations
Transparency Gaps in Account Deactivations
Foodinho used standardized messages during deactivations. These automated emails lacked specific reasons. Riders were left without clear contestation rights. This practice violates basic labor dignity.
Companies must ensure that automated systems do not lead to AI bias in hiring and recruitment discrimination. Opaque decision-making creates significant legal risks.
| HR Risk | Algorithmic Practice | Regulatory Requirement |
| Performance Rating | Hidden scores | Human intervention |
| Account Deactivation | Standardized emails | Right to contest |
| Geolocation Tracking | Background processing | Strict necessity |
| Data Access | Third-party sharing | Prior notification |
Transparency is a fundamental necessity. Employers must explain the logic behind every automated decision affecting a worker’s livelihood.
Foodinho and Glovo – From Privacy Enforcement to Labour Supervision
This recent €5 million fine is not a standalone event but rather the latest chapter in a long history of regulatory friction for the Glovo group.
The 2021 Precedent Involving Rider Ranking Systems
The 2024 fine is closely linked to a previous €2.6 million sanction from July 2021. The Garante identified recurring issues in how the platform manages its fleet. The core problems remain strikingly similar to those found in the 2021 Foodinho algorithmic fine details.
Rider-ranking algorithms have faced persistent accuracy concerns for years. The group failed to implement sufficient corrective measures following the initial investigation. This lack of transparency continues to penalize workers unfairly.
According to compliance statistics for 2026, regulatory pressure is mounting. Repeated violations often lead to exponentially higher fines and stricter oversight from national authorities.
These algorithms frequently fail to distinguish between legitimate absences and poor performance. Such technical rigidity leads to systemic discrimination against riders. It ignores basic worker protections like the right to strike or sick leave.
Collective Responsibility Within the Glovo Corporate Structure
Foodinho operates as a subsidiary of the Glovo group and was acquired by Delivery Hero in 2022. Corporate ownership does not shield subsidiaries from local labor laws. Regulatory accountability remains localized despite centralized technology.
We see a broader pattern of enforcement emerging across Europe. Spain and Italy have been particularly active in challenging gig economy models. The group now faces a coordinated regulatory front from multiple data protection authorities.
Maintaining global HR compliance in 2026 requires rigorous internal checks. Companies must learn how to conduct an HR audit in a multi-country setup to avoid these pitfalls.
Centralized algorithms require localized legal vetting to ensure compliance with specific national statutes. Failure to do so creates massive financial liabilities for the parent company. Legal risks cannot be outsourced to code.
What EU Platform Work and AI Rules Mean for Foodinho
Looking ahead, the regulatory pressure on Foodinho will only intensify as new European frameworks move from theory to enforcement.
The Platform Work Directive and Employment Status
Directive (EU) 2024/2831 on improving working conditions in platform work must be transposed into national law by 2 December 2026. It introduces rules on algorithmic management, transparency, human oversight and the correct determination of employment status for people performing platform work.
The Directive requires Member States to establish an effective, rebuttable legal presumption of an employment relationship where facts indicating direction and control are found under the relevant national law, collective agreements or practice, taking account of Court of Justice case law. The specific criteria and procedures that activate the presumption must be set out nationally. It is therefore not an automatic EU-wide reclassification of every platform worker.
Where the presumption applies, the platform must be able to rebut it by demonstrating that the contractual relationship is not an employment relationship under the applicable national definition. The Directive also seeks to make it easier for workers and their representatives to challenge incorrect classification without increasing their procedural burden.
For Foodinho, these rules are particularly relevant because the 2026 proceedings concern allegations that riders were formally treated as self-employed despite the degree of direction and control exercised through the platform. The outcome will depend on the applicable Italian law and the evidence in the proceedings, but the case illustrates the importance of assessing how work is actually organized rather than relying on the contractual label alone.
The Directive also establishes protections relating to algorithmic management. These include transparency about automated systems, human monitoring of decisions affecting platform workers, opportunities to obtain explanations and review, and restrictions on processing certain categories of personal data. Platforms should assess how their systems assign work, monitor performance, determine remuneration, restrict accounts and make decisions that materially affect workers.
Companies should prepare by mapping their algorithmic systems, documenting decision criteria, identifying decisions requiring human review and ensuring that workers can contest outcomes. They should also review contracts and operational practices against the national rules adopted to implement the Directive. The EU framework sets common requirements, but national legislation will determine important details of how those protections operate in practice.
AI Act Requirements for Employment and Worker Management
The EU AI Act provides a separate framework for certain AI systems used in employment and worker management. Systems used for recruitment, selection, decisions affecting employment terms, promotion or termination, task allocation based on individual behaviour or personal characteristics, and monitoring or evaluation of workers may fall within the high-risk category set out in Annex III, subject to the Regulation’s classification rules and exceptions.
The current application timeline has changed. Following the 2026 amendment to the AI Act, the relevant requirements for high-risk systems classified under Annex III are scheduled to apply from 2 December 2027, rather than the previously stated August 2026 date. This later application date does not suspend existing obligations under the GDPR or other applicable employment and data protection laws.
Foodinho’s case is primarily a GDPR enforcement matter, and not every algorithm used by a delivery platform automatically qualifies as a high-risk AI system under the AI Act. Nevertheless, platforms should assess whether tools used to rate workers, allocate tasks, monitor performance or influence access to work fall within the Act’s scope. Where they do, providers and deployers will need to address applicable requirements such as risk management, documentation, data governance, human oversight and monitoring.
Companies should treat the GDPR, the Platform Work Directive and the AI Act as complementary frameworks. Compliance requires more than adding a human reviewer to an automated process: organizations should be able to explain how decisions are made, provide meaningful routes to challenge them, monitor potential discriminatory effects and ensure that workers’ rights are respected throughout the process.
Wrapping Up
Foodinho’s case shows how algorithmic management can create overlapping risks involving data protection, worker rights, pay and employment status. The Garante’s €5 million fine in 2024 addressed unlawful data processing and inadequate safeguards around automated decisions. The 2026 labour proceedings brought additional scrutiny of rider compensation and alleged misclassification, while proposed pay changes remained under review.
Platforms should audit how their systems collect location data, rank workers, assign deliveries, calculate remuneration and restrict accounts. They should provide meaningful human review, explain decisions clearly and ensure that their contractual arrangements reflect the applicable legal requirements. Preparing for the Platform Work Directive’s 2 December 2026 transposition deadline and the AI Act’s updated high-risk application timeline will help organizations identify gaps before further enforcement action.













