Why does GDPR see personal data where you see none?
You anonymized the obvious fields, so you assume your customer data is no longer personal. Under GDPR, that assumption is usually wrong, and it is the mistake regulators price with fines.
Two definitions, one expensive gap
In the United States, PII means information that identifies a person, and the term itself has four common variants. Per Wikipedia's PII entry, these are not equivalent, and the effective definition shifts by jurisdiction and purpose. Under EU and UK regimes built on the GDPR, the term "personal data" is significantly broader. So a dataset that feels anonymous in a US engineering review can still be regulated personal data in Europe.
- US PII: information related to an identifiable person.
- GDPR personal data: broader scope, covering anything relating to an identifiable person.
Why this bites AI products hardest
Improper or missing disclosure control is often the root cause of privacy issues. AI tools collect, store, use, and finally destroy data, and privacy concerns exist at every one of those stages. Wikipedia's information privacy entry notes that improper or missing disclosure control is often the root cause of privacy issues. When your product pipes user inputs to a vendor, each hop inherits the privacy question. Free-text prompts are especially risky because people naturally type identifying details.
Consent is the mechanism you keep skipping
Informed consent mechanisms, including privacy policies and dynamic consent, are how you communicate data practices. Informed consent mechanisms, including privacy policies and dynamic consent, are how you communicate data practices, according to Wikipedia's privacy overview. The GDPR site organizes consent, privacy by design, and impact assessments as its key issues, not optional extras. That means data protection is a lifecycle you must design for, not a checkbox at signup. Most AI data-sharing deals fail quietly because consent was collected for one use and stretched to another.
What this does not solve
These sources define the rules; they do not tell you how to apply them to a specific vendor chain. The GDPR index lists articles and principles, not AI-era case law about model training or third-party processors. You still need legal review to map your actual data flows onto these definitions. Definitions alone will not tell you whether your logs are personal data in your specific case.
Which fields in your AI product's logs would a European regulator call personal data?
Was your consent collected for the exact data sharing your AI vendors now do?
List every third party your product sends user text to, then mark which ones your privacy policy actually names.
Sources
Wikipedia: Information privacy
Recommended for you
- Data
Why do your dashboards lie while your pipelines tell the truth?
When a metric becomes a target, people optimize the number, not the reality. Your raw pipeline has no incentive to flatter you.
- Data
Why does data quality belong to the business, not the data team?
Your data team can check schemas, but only the business knows what the data is for. Quality is fitness for purpose, and purpose lives outside engineering.
- Data
When should you store raw data instead of cleaning it first?
Your data choice locks in who can use it. Here is what the definitions alone tell you about warehouse versus lake.
- Data
Why does hiring one AI engineer still leave your data broken?
You hired for AI and still got outputs that contradict your dashboards. The problem is the role boundary nobody drew between data work and AI work.
Enjoyed this article?
Subscribe for new articles. No spam. Unsubscribe anytime.