When Algorithms Say No: What the €825 Million Uber Fine Teaches Us
- Author: Arno Schlösser, DP-Dock GmbH
- Last updated: November 2026
- Category: Enforcement, General Obligations
About AI and Data Privacy
Algorithms now decide who gets a loan, a job interview or the next ride. Under the GDPR, these practices fall under some of its most complex and possibly demanding rules: those on profiling and automated decision-making. With AI tools and data sharing now part of everyday business, it is a good moment to check whether your organization profiles, and whether it does so lawfully.
Profiling is the automated use of personal data to evaluate or predict aspects of a person, such as their performance, financial situation, preferences, behavior or location. It is not strictly prohibited, but it must comply with all GDPR principles – for example, companies must be transparent about it, and individuals have the right to object to it.
Automated decision-making goes beyond this. Individuals have the right not to be subject to decisions based solely on automated processing that have legal or similarly significant effects, such as personalized rates on services or products. Such decisions are allowed only on limited grounds and with safeguards, including the right to human intervention. Individuals must also be informed about them as transparently as possible and receive meaningful information about the logic involved.
The Court of Justice of the EU has recently raised the bar. In SCHUFA (2023), it held that a credit score can itself be an automated decision. In Dun & Bradstreet Austria (2025), it confirmed that explanations must be genuinely understandable. And a human who merely rubber-stamps an algorithmic outcome does not take the decision outside Article 22.
Why the Whole Issue Matters More Than Ever
AI systems and algorithms evaluate and predict human behavior by design. Recruitment tools, fraud models, pricing engines and customer analytics all profile, often through off-the-shelf software whose “smart features” quietly bring organizations within scope of these rules. At the same time, complex models are hard to explain, which makes meeting the transparency requirements more difficult.
Data sharing adds to the risk. Profiles are increasingly built by combining data from partners, group companies and data brokers, and every data flow needs a lawful basis – all the more so as increasing globalization means that data often leaves the EEA. Meanwhile, the EU AI Act adds its own obligations for high-risk uses such as recruitment and credit scoring, on top of the GDPR. With fines of up to 4% of worldwide annual turnover, the stakes are high.
The Issue Attracts Growing Regulatory Attention
On 8 October 2026, the Dutch Data Protection Authority (AP) fined Uber €824,990,000 for unlawful automated decision-making and insufficient information provided about profiling its drivers. Uber automatically deactivated drivers’ accounts: temporarily in case of suspected fraud or customer ratings that were too low, and permanently in case of persistently low ratings. During a deactivation, drivers lost their income via Uber. The AP found that these were fully automated decisions in breach of Art. 22 GDPR, and that Uber had not sufficiently informed drivers about them as required by Art. 13 GDPR.
The investigation started with complaints from almost 200 French drivers, submitted to the French supervisory authority, the CNIL. Because Uber’s European headquarters is in the Netherlands, the AP handled the case. The incidents took place between 2018 and 2022. Uber has since stopped the infringements and has appealed the fine; there is no final judicial decision yet.
In any case, the regulatory action shows that decisions with significant implications for individuals must involve meaningful human oversight and transparent communication – and that regulators are ready to impose substantial fines where these requirements are not met.
Do You Profile?
Many organizations profile without realizing it. Activities recorded as “CRM”, “analytics” or “HR administration” may in fact involve evaluating or predicting people’s behavior, often through AI features in third-party tools. We therefore recommend revisiting your processing activities with a fresh eye to understand whether and where profiling takes place and whether any significant decisions are made without adequate human involvement. Where that is the case, make sure it meets all requirements of the GDPR.
With AI embedded in everyday tools and data flowing across organizations and borders, profiling is no longer a niche topic. Knowing where and how you profile is the first step to staying compliant. If you would like support with this review, please get in touch.