Privacy in focus: What’s in a word?

In this post from our ‘Privacy in focus’ blog series, we explore arguments for and against changes to the definition of personal information being considered by the review of the Privacy Act, and the implications of those changes.

One of the simplest but most far-reaching potential amendments to the Privacy Act is the replacement of a single word: replacing ‘about’ with ‘relates to’ in the definition of ‘personal information’.

Supporters of the change (such as the ACCC, the OAIC, and the Law Council of Australia) say it would clarify significant legal uncertainty, while also aligning Australia with the GDPR standard and maintaining consistency between the Privacy Act and the Consumer Data Right regime.

Those opposed (such as the Communications Alliance and the Australian Industry Group) warn that the change may unnecessarily broaden the scope of the Act, potentially imposing substantial costs on industry without any clear benefit to consumers.

To understand why, we’ll dig into the origins of the definition and the present uncertainty regarding its application.

Precision is important

The definition of personal information sets the scope of the Privacy Act. All the rights and obligations in the Act rely on this definition. All the obligations that organisations have to handle personal information responsibly rely on this definition. All the rights that individuals have to control how their personal information is used rely on this definition.  Personal information is the very base on which privacy regulation rests.

Any uncertainty in such an important definition can result in significant costs for both individuals and organisations. At best, uncertainty can result in wasted compliance work governing and controlling data that need not be protected. At worst, it can mean severe violations of privacy for consumers when data breaches occur as a result of failure to apply controls to data that should have been protected. Examples of the former are frequent — even OAIC guidance encourages organisations to err on the side of caution in identifying data as personal information. Unfortunately, examples of the latter are even more commonplace — the disclosure of Myki travel data by Public Transport Victoria, the publication of MBS/PBS data by the Federal Department of Health, and Flight Centre’s release of customer data for a hackathon are all recent examples of organisations releasing data subject to inadequate controls in the belief that it did not amount to personal information.

These uncertain times

According to the OAIC, the ACCC, and many others, there is substantial uncertainty as to the scope of ‘personal information’, particularly as it relates to metadata such as IP addresses and other technical information. That uncertainty was partially created, and certainly enhanced, by the decision of the Administrative Appeal Tribunal in the Grubb case, which was upheld on appeal in the Federal Court.

In the Grubb case, the Tribunal found that certain telecommunications metadata was not personal information because it was really ‘about’ the way data flows through Telstra’s network in order to deliver a call or message, rather than about Mr Grubb himself.

The ruling came as a surprise to many. The orthodoxy up until that point had been that the word ‘about’ played a minimal role in the definition of personal information, and that the relevant test was simply whether the information is connected or related to an individual in a way that reveals or conveys something about them, even where the information may be several steps removed from the individual.

Today, it’s still unclear how significant a role ‘about’ should play in the definition. Could one argue, for example, that location data from a mobile phone is information about the phone, not its owner? Or that web browsing history is information about data flows and connections between computers, rather than about the individual at the keyboard?

OAIC guidance is some help, but it’s not legally binding. In the absence of further consideration by the courts, which is unlikely to happen any time soon[1], the matter remains unsettled. Organisations are without a clear answer as to whether (or in what circumstances) technical data should be treated as personal, forcing them to roll the dice in an area that should be precisely defined. Individuals are put in the equally uncertain position of not knowing what information will be protected, and how far to trust organisations who may be trying to do the right thing.  

Relating to uncertainty

Those in favour of reform want to resolve this uncertainty by replacing ‘about’ with ‘relates to’. The effect would be to sidestep the Grubb judgement and lock in a broad understanding of what personal information entails, so that the definition covers (and the Privacy Act protects) all information that reveals or conveys something about an individual, including device or technical data that may be generated at a remove.

Those who prefer the status quo take the view the present level of uncertainty is manageable, and that revising the definition to something new and untested in Australia may lead to more confusion rather than less. Additionally, there is concern that ‘relates to’ may represent a broader test, and that the change could mean a significant expansion of the scope of the Act into technical and operational data sets.

What we think

By drawing attention to ‘about’ as a separate test, the Grubb case has led to an unfortunate focus on how information is generated and its proximity to an individual, when the key concern of privacy should always be what is revealed or conveyed about a person. In our view, replacing ‘about’ with ‘relates to’ better focuses consideration on whether an identifiable individual may be affected.

Industry concerns about expanding the scope of the Act are reasonable, particularly in the telco space, though we anticipate this to be modest and manageable as the scope of personal information will always remain bounded by the primary requirement that personal information be linked back to an identifiable individual. Further, we anticipate that any additional compliance costs will be offset by a clearer test and better alignment with the Consumer Data Right and Telecommunications (Interception and Access) Act, both of which use ‘relates to’ in defining personal information.

Finally and significantly for any businesses operating outside of Australia, amending ‘about’ to ‘relates to’ would align the Privacy Act more closely with GDPR. Aligning with GDPR will be something of a recurring theme in any discussions about the Privacy Act review. This is for two reasons:

  • GDPR is an attractive standard. GDPR has come to represent the de-facto global standard with which many Australian and most international enterprises already comply. It’s far from perfect, and there are plenty of adaptations we might want to make for an Australian environment, but generally aligning to that standard could achieve a high level of privacy protection while minimising additional compliance costs for business.
  • Alignment might lead to ‘adequacy’. The GDPR imposes fewer requirements on data transfers to jurisdictions that the EU determine to have ‘adequate’ privacy laws. A determination of adequacy would substantially lower transaction and compliance costs for Australian companies doing business with the EU.

Click ‘I agree’ to continue

In our next edition of the Privacy in Focus series, we’ll take a look at consent and the role it might play in a revised Privacy Act. Will Australia double down on privacy self-management, or join the global trend towards greater organisational accountability?

Footnote: [1] Because of the way that privacy complaints work, disputes about the Privacy Act very rarely make it before the courts — a fact we’ll dig into more when we cover the proposal for a direct right of action under the Act.


Read all posts from the Privacy in focus series:
Privacy in focus: A new beginning
Privacy in focus: Who’s in the room?
Privacy in focus: What’s in a word?

Privacy in focus: Who’s in the room?

In this post from our ‘Privacy in focus’ blog series, we explore the key voices and perspectives shaping the review of the Privacy Act.

If you want to know where the review of the Privacy Act is going to land, the first question to ask is ‘who’s in the room’.

That’s why, in this post on the Privacy Act review, we’ve analysed public submissions in response to the Government’s issues paper to see what they reveal about the perspectives of interest groups, and how this might shape the review process.

It’s loud in here

There are 154 submissions published on the Attorney General’s website, totalling 2,103 pages by our count. That’s quite a few by comparison with other consultation processes. The ACCC’s Digital Platforms Inquiry issues paper only attracted 76 submissions.

More than half of all submissions come from private companies (around 30%) and industry bodies or professional associations (around 23%). Within this segment, a wide range of industries are represented – it really is a cross section of the economy. Contributions from the Shopping Centre Council of Australia, the Obesity Policy Coalition and the Federal Chamber of Automotive Industries might have been surprising a few years ago. Today their presence is a testament to how central data has become in our lives.

The remaining submissions come from academics and research centres (around 16%), various government entities (around 13%), charities and advocacy groups (around 10%) and individuals (around 7%).

Reading the room

There are so many issues and so many differing interests and perspectives that it is difficult to draw many clear through-lines. By our rough (and inevitably subjective) count:

  • A little over 50% of all submissions are broadly in support of stronger privacy protections.
  • Around 20% advocate little or no change to the current regime.
  • The remainder are either explicitly neutral, focus on a specific issue or provide commentary on a specific industry without taking a clear position.
  • Only a small handful of submissions advocate for weaker protections.

What’s the chatter?

The small business and employee records exemptions are shaping up as a key battleground, with an unlikely alliance between privacy advocates (Electronic Frontiers Australia, New South Wales Council for Civil Liberties) and tech/data companies (Google, Data Republic) against the exemptions on one side, pitted against representatives of small business and sole traders in a range of fields (Arts Law Centre of Australia, Clubs Australia and the Australian Small Business and Family Enterprise Ombudsman) favouring the exemption on the other.

The role of consent will be another area of contention. A large number of submissions have raised concerns about the ACCC Digital Platforms Inquiry recommendations for enhanced consent requirements. Some note the failure of the notice and consent model as a whole and emphasise the need for additional controls on how organisations use data (see particularly the Consumer Policy Research Centre and the Association for data-driven marketing and advertising). Others emphasise the dangers of consent fatigue and the need for an alternative basis for processing (see e.g., Facebook).

Finding your friends – opposing unnecessary regulation

As one might expect, submissions from industry are more likely to oppose or raise concerns about higher regulatory standards. Those worried about the potential costs of reform include:

Finding your friends – supporting higher standards

Perhaps surprisingly, many of the most data centric businesses and industry groups support reform. Data service providers (such as illion and Experian), advertisers (such as the Association for data-driven marketing and advertising), and software/technology services (such as Adobe, Atlassian, Data Republic) are much more open to reform, particularly in pursuit of GDPR adequacy.

Submissions from human rights groups (such as Digital Rights Watch, New South Wales Council for Civil Liberties) and consumer advocacy groups (such as Australian Communications Consumer Action Network, CHOICE, Financial Rights Legal Centre, Consumer Action Law Centre and Financial Counselling Australia) near-universally support greater protections, as do academics (such as the Centre for AI and Digital Ethics and Melbourne Law School, University of Melbourne, Castan Centre for Human Rights Law – Monash University) and professionals (such as Australian Information Security Association and the Law Council of Australia) also skew heavily towards stronger privacy rules.

What next?

Our takeaway is that there are substantially more voices in favour of reform than for the status quo. Add that to the overwhelming public support for stronger privacy protections (83% of Australians surveyed by OAIC saying they would like the government to do more to protect the privacy of their data) and it looks like there will be real pressure on the government to deliver meaningful reform.

Of course, the issues paper is just the beginning, and we’ve just scratched the surface here. So why not stay tuned while we all wait for the discussion paper? In our next edition, we’ll take a deep dive into the definition of personal information.


Read all posts from the Privacy in focus series:
Privacy in focus: A new beginning
Privacy in focus: Who’s in the room?
Privacy in focus: What’s in a word?

When your milk comes with a free iris scan

elevenM’s Melanie Marks’ regular trip to the supermarket brings her face-to-face with emerging privacy issues.

A couple of weeks ago, as I was nonchalantly scanning my groceries, I looked up and was shocked to see a masked face staring back at me. 

After I realised it was my own face, fright turned to relief and then dismay as it hit me that the supermarkets had – without consultation, and with limited transparency – taken away my freedom to be an anonymous shopper buying milk on a Sunday.

Just days later, the press outed Coles for its introduction of cameras at self-service checkouts. Coles justified its roll-out on the basis that previous efforts to deter theft, such as signs that display images of CCTV cameras, threats to prosecute offenders, bag checks, checkout weighing plates and electronic security gates have not been effective and the next frontier is a very close-up video selfie to enjoy as you scan your goodies.

Smart Company reported on the introduction of self-surveillance tech last year, explaining the psychology of surveillance as a deterrent against theft. How much a person steals comes down to their own “deviance threshold” — the point at which they can no longer justify their behaviour alongside a self-perception as a good person.

The supermarkets’ strategy of self-surveillance provides a reminder that we are being watched, which supposedly evokes self-reflection and self-regulation.

This all sounds reason enough. Who can argue with the notion that theft is bad, and we must act to prevent it? We might also recognise the supermarkets’ business process excellence in extending self-service to policing.

Coles argues that they provide notice of the surveillance via large posters and signs at the front of stores. They say that the cameras are not recording, and they claim that the collection of this footage (what collection – if no record is being made?) is within the bounds of its privacy policy (last updated November 2018).

At the time of writing this blog, the Coles privacy policy makes no mention of video surveillance or the capturing of images, though it does cover its use of personal information for “investigative, fraud, and loss prevention” activities.

Woolworths has also attracted criticism over its use of the same software, which it began trialling last year. Recent backlash came after Twitter user @sallyrugg called on the supermarket to please explain any connection between the cameras, credit card data and facial recognition technology it employs. Like Coles, Woolies says no recording takes place at the self-serve registers and that the recent addition it has made to its privacy policy regarding its use of cameras pertains only to the use of standard CCTV in stores.

So it would appear the supermarkets have addressed the concerns. No recordings, no data matching, covered by privacy policy. And my personal favourite: choice: “If you do not wish to be a part of the trial, you are welcome to use the staffed checkouts.

But these responses are not sufficient. Firstly, there is no real choice in relation to the cameras when a staffed checkout is unavailable. Secondly, our notice and consent models are broken, which overstates the actual power granted to consumers by privacy policy. We don’t read them, and even when we do, we have no bargaining power. And lastly, the likelihood of function creep is high. It is not a stretch to imagine that the next step in the trial will be to pilot the recording of images for various purposes, and it could be navigated legally with little constraint.

On a final note, this experience reflects many of the challenges in our current privacy framework including: the balance of consumer interests against commercial interests, the strain on current consent models, and even the desire for a right to be forgotten.

Thankfully, these issues are all being contemplated by the current review of the Privacy Act (read our ongoing blog series on the review here). We need these protections and structures in place, to create a future in which we milk buyers can be free and anonymoos.

Photo by Ali Yahya on Unsplash