Interview: Give staff access to info on a need-to-know basis
Australian Privacy Foundation chair discusses the Vodafone security breach
In the wake of the mass breach of Vodafone security last week, The Register spoke to Roger Clarke, chair of the Australian Privacy Foundation and long-time privacy campaigner, about what the industry can learn from the incident. Staff and partners at Vodafone have logins that allow them access to customer data on that database – and these details were leaked.
El Reg: The security breach at Vodafone has been described as “customer details being published on the Internet”. We’ve said that Vodafone did not “publish” information – its intention was that the information was meant to remain private. You’ve indicated that you believe the issue to be more serious than that – that we might be understating the issue. So that looks like a good place to start!
Clarke: Let me go back a step. The idea that the Web can be used as a means of achieving sufficiently secure extranet and even Intranet applications is fine. I don’t have a problem with the idea that Vodafone is doing that.
There are potential concerns that aren’t specific to Vodafone. One is that when you provide anyone with access to a database, you must provide them with access only to the data they need to access to perform their functions.
An example of access to too much data may be that most members of Vodafone staff don’t need access to credit card numbers. They may need access to the fact that people HAVE credit card details registered with the company, but not the number.
Likewise, the majority of staff members of Vodafone have no need to get access to call records.
Finally, the idea that content – for example, the contents of SMSs – may be accessible by staff members is more serious. And recording calls would, of course, break interception laws.
Once you go to resellers, you’re losing the organisational control over your employees. The question is: “What do you need to have access to?”
In practical terms, we may have to allow a reseller in Broome to assert that they have Customer X in front of them and make an exception by permitting access by someone other than the customer’s usual or previous re-seller – but then you would want to log it!
There are accusations that there have been stuff-ups by Vodafone in that regard – either access to large amounts of data about a customer, or data about lots of customers.
The other concern is access by someone outside the organisation who the a staff-member or re-seller passes their login details to – anything the staff member has access to is then accessible to a the party outside, either by connivance or purchase.
This sort of risk is normally controlled in a number of ways; unauthorised access is limited to only that data to which the individual has access, and access to data about lots of customers is limited to a small number of staff. You will grant more access to your security people, but they will be vetted more closely.
El Reg: None of these issues are specific to Vodafone, though. Anyone using a similar portal would have to deal with them.
Clarke: That risk always exists. That’s why you have controls and audits of the controls.
El Reg: What does the Privacy Act have to say about “bad apple” incidents – where the company didn’t intend to make customer data public, but an insider has breached its policies?
Clarke: The Privacy Act is almost irrelevant. It’s toothless and administered by an organisation that doesn’t spend much time exercising those powers it has.
The problem isn’t the obligations, it’s the unenforceability of the obligations. [The Act] requires that security precautions be taken – and it’s covered by a reasonableness clause, so that you don’t have to have extraordinarily expensive security on not-very-interesting data.
There is an obligation, but what hasn’t happened is for the Privacy Commissioner to get out there, and get consultancies to advise it on how those obligations should be implemented in particular circumstances.
El Reg: But the Privacy Commissioner built a much higher profile last year in the Google WiFi “sniffing” controversy. Doesn’t that indicate that the office is becoming more active?
Clarke: The PC did an investigation – and they said “Google agreed it did something wrong and agreed that they won’t do it again”.
But because it was under the Commissioner’s investigation powers, it doesn’t have to publish what it found in its investigation. It was spineless and completely useless – what is the point of a regulator if they don’t regulate?
The Australian Federal Police decided not to prosecute, and we [the Privacy Foundation] have to do an FOI request to find out why.
There should have been significant undertakings, and reparations of a monetary nature that could be used for an appropriate purpose – but nothing was done.
And then there’s data breach notification. There’s been four to five years of movement in the area of data breach notification, and we have very little to show for that. There’s been a report that’s gone nowhere, there are Law Reform Commission proposals, there’s a non-binding draft by the Privacy Commissioner, but there’s nothing in [Commonwealth] law.
Vodafone has done one thing, which is to send out a letter to subscribers – whether all subscribers, or only those that may have been affected I don’t know – but it doesn’t go very far, because Vodafone was still investigating what had happened.
The public needs information.
El Reg: Isn’t that the law at fault? Vodafone could argue that it’s complying with the law.
Clarke: Yes, possibly. But there’s also a public policy position that’s generally accepted by organisations. They have to do more than just stick with the minimum set by law; they have to assess the reasonable interests of the public.
You have to judge what is appropriate in the circumstances – companies have an obligation to provide sufficient information to the public, and the Commissioner has to ensure that happens.
We need enough information out there so that other companies can assess their risk – get their security officer to run a test on their own systems to see if the inadequacies exist in their own systems. ®
Sponsored: Today’s most dangerous security threats