Dutch Regulator Fines Uber €825m Over Automated Suspensions
A landmark GDPR automated-decision penalty reinforces that absence of meaningful human intervention in adverse outcome decisions is a core governance risk - directly relevant to APS ADM frameworks.
Key points
- Dutch and French regulators fined Uber €824.99 million for fully automated driver account deactivations under GDPR.
- Regulators found no human intervention in decisions that could remove drivers' ability to earn income - the decisive issue.
- Australian agencies using automated decision-making that affects individuals' access to services face analogous governance questions.
Implications for Australian agencies
- Consider Agencies using automated or algorithm-assisted decisions that can restrict access to services or payments may want to assess whether their human review processes are substantive and documented rather than nominal.
- Consider Teams designing or procuring AI-assisted enforcement or eligibility systems could use this case's regulatory reasoning to stress-test their own notice, review, and appeal controls against existing Australian Government ADM obligations.
Implications are AI-generated. Starting points, not advice — see methodology for how they're framed.
View original source
Copied.
"Dutch Regulator Fines Uber €825m Over Automated Suspensions"
Source: Let's Data Science – AI Governance
Published: 25 August 2026
URL: https://letsdatascience.com/news/dutch-regulator-fines-uber-over-automated-suspensions-536aa002
Dutch and French privacy regulators have jointly fined Uber €824.99 million for GDPR violations covering automated temporary and permanent deactivations of driver accounts between 2018 and 2022. The decisive regulatory finding was the complete absence of human intervention in decisions that could materially prevent drivers from earning income. Uber disputes the ruling and will appeal, noting current policies include human review and appeal mechanisms. For APS practitioners, the case illustrates that regulators assess automated decision-making not solely on model accuracy but on the governance architecture surrounding it - specifically whether human review is substantive, appealable, and capable of changing the outcome.
Implications for Australian agencies:
- [Consider] Agencies using automated or algorithm-assisted decisions that can restrict access to services or payments may want to assess whether their human review processes are substantive and documented rather than nominal.
- [Consider] Teams designing or procuring AI-assisted enforcement or eligibility systems could use this case's regulatory reasoning to stress-test their own notice, review, and appeal controls against existing Australian Government ADM obligations.
Retrieved from SIMS, 16 September 2026.