Lockstep

Mobile: +61 (0) 414 488 851
Email: swilson@lockstep.com.au

Google's wifi misadventure, and the gulf between IT and Privacy

An article for the iappANZ, examining what Google's carelessness with data collection tells us about technologists' appreciation of privacy.

See also Public yet still private.

Google’s wifi misadventure, and the gulf between IT and Privacy


Everyone knows by now that privacy and security are not the same thing; control is the name of the game. Yet many responses to Google’s most recent misadventure over wireless network data show we still have a long way to go. Too many are naively downplaying the incident as if privacy and secrecy are equivalent.

It’s understandable for people to have a range of views about privacy. Not only is it personally felt, it is also a soft and multi-disciplined field, involving philosophy, human rights, civil liberties and politics. Yet information privacy law turns out to be rather clinical and is written to neatly sidestep philosophical and moral mine fields. This should make privacy law accessible to a wider audience, and yet its implications for IT remain misunderstood by many.

A few weeks ago it came to attention that Google Street View cars were collecting information from wireless networks they drove by. The extent of this collection wasn’t clear at first, and speculation was confined to the seriousness of recording network names and device addresses transmitted publicly by wifi installations. In itself, basic network data is not identifying, but a nice privacy question arises if it can be linked to named individuals through Google’s other vast data sets.

Google then revealed that their cars were also sampling regular data traversing unencrypted networks within range. Google explained that engineering software for an experimental wifi project was inadvertently included in the Street View system.

[It appears to me that a developer working on the system took an interest in the Wi-Fi data and was simply curious about what was out there and available to be sampled. So they added some more code that saved to the hard drive any un-encrypted traffic collected while the network was in range of the car. Presumably they planned to explore the data later, and see what they could see. Even if the motivation was ostensibly harmless curiosity, what's remarkable that anyone feels that they can help themselves to information just because it's there. I suspect that what runs deep in Google is a thirst for data and a drive to use whatever they can.]

On blogs and discussion groups, many downplayed the collection of “public” data. Some blamed the victims, insisting that if network operators haven’t taken care to turn on encryption, then their “broadcasts” are up for grabs. Yet carelessness does not equate to granting blanket consent for personal information to be collected and used for anything a technology company thinks of. Privacy law, like the Trade Practices Act, serves to protect the ‘little guy’ against exploitation by big business.

The critical legal point to me is crystal clear: information privacy law prohibits businesses from collecting personal information without an express need, or without consent, whether the source data is in the public domain or not.

When they learn about how the Privacy Act treats collection, many IT practitioners are incredulous, pointing out that vast volumes of information are generated and shared as a by-product of how networks work. And this fact of digital life does need to be understood by lawyers, legislators, policy makers, technologists and auditors.
Technical breaches of the Privacy Act probably occur all the time, by virtue of how computer systems operate, and in particular, how they tend to generate and hang on to log data.

A classic example is the personal data logged by e-commerce servers. Peoples’ shopping habits represent personal information and e-merchants may fall foul of privacy law if they are logging transaction details without an express need. Worse, if the purchasing history includes medicines or even herbal remedies, then e-merchants are likely collecting Sensitive Information. Naturally customers’ buying patterns can be an important marketing resource, but as such it can only be legally exploited if collected with consent and the primary purpose (i.e. marketing) has been made clear to all concerned.

Many organisations may find themselves in trouble if their technicians have enabled rich logging because they guess the information might be useful one day, without being specific and open about it. The engineer at Google appears not to have understood the Collection Principle, and I’m guessing that the culture at Google has not internalised the privacy principles in general.

Technologists especially should heed several lessons from the wifi misadventure:

  • It might be counter-intuitive but publicly available information is still subject to information privacy law if it relates to identifiable individuals.
  • Terminology matters (as technicians know in their own fields). The words “public” and “private” are not precise, so privacy law avoids them and instead is very specific about personal information.
  • Engineers must resist the temptation to collect personal information just because they can.
  • And software development processes should include checks to see if excessive information collection is going on for whatever reason before applications are released into the wild.