A famous experiment was when Yahoo (I think) released a dump of anonymised search queries, however, they had a user identifier for each query so it was simple enough to combine the queries and find the person.

It was AOL, but yea, they didn't really think their method of "anonymizing the data" through.

The definition of PII needs to be expanded to "combinations of attributes that can uniquely identify a single individual with high probability."

For the EU, the definition in the GDPR should cover this. From Article 4(1):

> ‘personal data’ means any information relating to an identified or identifiable natural person (‘data subject’); an identifiable natural person is one who can be identified, directly or indirectly, in particular by reference to an identifier such as a name, an identification number, location data, an online identifier or to one or more factors specific to the physical, physiological, genetic, mental, economic, cultural or social identity of that natural person;

I'm a fun example here too. Through a quirk of how addresses are assigned in my city, and how postal codes are assigned in Canada, my house has a unique postal code. Given our postal code, you have 100% certainty, and 50% probability between me or my partner.

That’s already the definition used in many contexts.

What contexts are you thinking of? I’d agree in the context of discussing de-anonymizing data. But all the policy, e.g., GDPR, and licenses that I’ve seen so far mean specifically 100% directly identifiable (such as name & address or SSN) when they say “personally identifiable information”, and especially when they use the acronym PII.

No, that's not what GDPR says.

https://gdpr-info.eu/art-4-gdpr/

> ‘personal data’ means any information relating to an identified or identifiable natural person (‘data subject’); an identifiable natural person is one who can be identified, directly or indirectly, in particular by reference to an identifier such as a name, an identification number, location data, an online identifier or to one or more factors specific to the physical, physiological, genetic, mental, economic, cultural or social identity of that natural person;

[...]

> ‘profiling’ means any form of automated processing of personal data consisting of the use of personal data to evaluate certain personal aspects relating to a natural person, in particular to analyse or predict aspects concerning that natural person’s performance at work, economic situation, health, personal preferences, interests, reliability, behaviour, location or movements;

Yeah you’re right. I stand corrected, and I even read that section last week.

From GDPR Article 4(1):

> ‘personal data’ means any information relating to an identified or identifiable natural person (‘data subject’); an identifiable natural person is one who can be identified, directly or indirectly, in particular by reference to an identifier such as a name, an identification number, location data, an online identifier or to one or more factors specific to the physical, physiological, genetic, mental, economic, cultural or social identity of that natural person;

To layer on some interpretation: * Red Shell is collecting information * which relates to "an identifiable natural person" * who "can be identified [...] by reference to [...] one or more factors specific to the physical [...] identity of that natural person" * where these factors are the information Red Shell is collecting about the software and hardware of their physical devices * and thus are "personal data", as regulated by the GDPR.

Given that their primary purpose is identifying users, they're going to have a hard time arguing that they don't fall under the GDPR's jurisdiction.

One of the various "protecting personal information" trainings I had to do for a US Govt job made mention of multiple non-identifiable pieces of information that combined could identify a person falling under PII, but how many is ok to put together was left purposely ambiguous.

Just the fact that I browse exclusively in private mode from Australia, don't accept third party cookies, apparently have Ghostery installed and use Safari vX.Y is enough to narrow down my identity to about one person give or take.

The information collected in this program is more than enough to identify me personally, and probably enough to start making some kind of psychological profile to help them predict how to convince me to vote for Donald Trump (or in my case, a Liberal Party member).