Head in the Sandbox

The Information Commissioner’s Office recently held a workshop about their proposed Regulatory Sandbox. The idea of the sandbox is that organisations can come to the ICO with new proposals in order to test out their lawfulness in a safe environment. The hoped-for outcome is that products and services that are at the same time innovative and compliant will emerge.

There is no mention of a sandbox process in the GDPR or the DPA 2018. There is a formal mechanism for controllers to consult the ICO about new ideas that carry high risk (prior consultation) but the circumstances where that happens are prescribed. It’s more about managing risk than getting headlines. Unlike Data Protection Impact Assessments, prior consultation or certification, the design and operation of the sandbox is entirely within the ICO’s control. It is important to know who is having an influence its development, especially as the sandbox approach is not without risk.

Although Mrs Denham is not above eye-catching enforcement when it suits her, the ICO is often risk averse, and has shown little appetite for challenging business models. For example, the UK’s vibrant data broking market – which is fundamentally opaque and therefore unlawful – has rarely been challenged by Wilmslow, especially not the bigger players. They often get treated as stakeholders. The sandbox could make this worse – big organisations will come with their money-making wheezes, and it’s hard to imagine that ICO staff will want to tell them that they can’t do what they want. The sandbox could leave the ICO implicated, having approved or not prevented dodgy practices to avoid the awkwardness of saying no.

Even if you disagree with me about these risks, it’s surely a good thing that the ICO is transparent about who is having an influence on the process. So I made an FOI request to the ICO, requesting the names and companies or organisations of those who attended the meeting. As is tradition, they replied on the 20th working day to refuse to tell me. According to Wilmslow, disclosure of the attendees’ identities is exempt for four different reasons. Transparency will prejudice the ICO’s ability to carry out its regulatory functions, disclosure of the names of the attendees is a breach of data protection, revealing the names of the organisations will cause them commercial damage, and finally, the information was supplied with an expectation of confidentiality, and so disclosure will breach that duty.

These claims are outrageous. DPIAs and prior disclosure exist, underpinned both by the law and by European Data Protection Board guidance. Despite the obvious benefits of developing a formal GDPR certification process (both allowing controllers to have their processing assessed, and the creation of a new industry at a time when the UK needs all the economic activity it can get), the ICO’s position on certification is supremely arrogant: “The ICO has no plans to accredit certification bodies or carry out certification at this time“. A process set out in detail in the GDPR is shunned, with the ICO choosing instead to spend huge amounts of time and money on a pet project which has no legal basis. Certification could spread expertise across the UK; the sandbox will inevitably be limited to preferred stakeholders. If they’re hiding the identities of those who show up to the workshop, it’s hard to imagine that the actual process will be any more transparent.

The ICO’s arguments about commercial prejudice under S43 of FOI are amateurish: “To disclose that a company has sent delegates to the event may in itself indicate to the wider sector and therefore potential competitors that they are in development of, or in the planning stages of a new innovative product which involves personal data“. A vital principle of FOI is that when using a prejudice-based exemption, you need to show cause and effect. Disclosure will or will be likely to lead to the harm described. How on earth could a company lose money, or become less competitive, purely because it was revealed that they attended an ICO event (which is what using S43 means)?

The ICO’s personal data and confidentiality arguments are equally weak – everyone who attended the meeting would know the identities of everyone else, and all were acting in an official or commercial capacity. This was not a secret or private meeting about a specific project; anyone with an interest was able to apply to attend. Revealing their attendance is not unfair, and there is plainly a legitimate interest in knowing who the ICO is talking to about a project into which the office is putting significant resources, and which will have an impact on products or services that may affect millions of people. The determination to hide this basic information and avoid scrutiny of the sandbox process undermines the credibility of the project itself, and makes the ICO’s claim to be an effective defender of public sector transparency ever more hypocritical.

Worst of all, if disclosure of the attendees’ identity was the calamity for commercial sensitivity and personal data that the ICO claims it to be, there should be an immediate and thorough investigation of how the information I requested came to be revealed on the ICO’s website and twitter account. The entire event was recorded and a promotional video was released. Several attendees (whose names and companies I cannot be given because of confidentiality, data protection and commercial prejudice) are identified and interviewed on camera, while there are numerous shots of other attendees who are clearly identifiable. Either the ICO has betrayed the confidentiality and personal data rights of these people, putting their companies at direct commercial risk, or their FOI response is a cack-handed attempt to avoid legitimate scrutiny. Either way, I strongly recommend that the left hand and the right hand in Wilmslow make some rudimentary attempts to get to know one another.

Long ago, I was one of a number of online commentators described by the ICO’s comms people as a ‘driver of negative sentiment’. More recently, one of Denham’s more dedicated apologists accused me of being one of the regulator’s “adversaries”. I’m not a fan of the ICO, and I never have been. But this stinks. The determination to throw every conceivable exemption at a simple request to know who the ICO is talking to suggests that the office is afraid of scrutiny, afraid of having to justify what they’re doing and how they’re doing it. The incompetence of refusing to give me information that is on display on their website and Twitter account shows contempt for their obligations as an FOI regulator. The ICO has its head in the sand; as we drift out of the European mainstream into a lonely future on the fringes, their secrecy and incompetence should be matters of concern for anyone who cares about Data Protection.

Bad Policy

On July 19th 2018, Linda McKee made a simple (but admirably polite) FOI request to the Information Commissioner’s Office. McKee asked for a copy of the ICO’s special categories policy document, a requirement of the Data Protection Act 2018 when processing special categories data in certain circumstances. The DPA was passed in early May 2018, but the requirement for special categories policies had been known since the DP Bill was published in September 2017. Policy documents were not required under the previous DP regime, and having run training courses on both the Bill and the Act, I can confirm that many people in the sector were keen to see real life examples of a policy document. McKee’s request made a lot of sense.

On 17 August (maintaining the ICO’s flawless record of replying to FOIs at the last minute), Wilmslow responded. They confirmed that a policy document was held, but as there was a clear intention to publish the policy document in the future, they refused to disclose it. This seemed a bit daft to me; Section 22 of FOI is designed to protect the organisation from early publication of information. The revelation of the ICO’s special categories policy would hardly cause ripples throughout the sector. Staff would not have been diverted from their normal jobs to deal with the torrent of press attention its release would provoke. They should have coughed it up and moved on.

McKee asked for an internal review, and at this point, the Commissioner headed determinedly the wrong way. There is no fixed time limit for an internal review, which is a flaw in the legislation but nevertheless not something that the organisation should exploit, and the ICO dragged it out for MONTHS. I have to be honest, I didn’t really pay attention, aside from using the ICO’s inability to release a relatively simple document as a gag on my DPA courses. Towards the end of 2018, I checked back in on McKee’s woes, to see an interesting suggestion on the What Do They Know thread. It seemed that when the ICO replied in August, the policy hadn’t actually been finalised.

I couldn’t quite believe this, so over Christmas, I made an FOI request to clear the matter up. I asked whether the policy was held in a final approved form when the ICO replied to McKee in August, for any recorded information about whether the ICO should actually have replied that the policy was not held (because it was not finished), and for a summary of why the ICO refused the request.

And here, a brief interlude to consider a section of the FOI Act that has tantalised FOI experts for years without resolution. Section 77 makes it a criminal offence for the organisation to alter, deface, block, erase, destroy or conceal any record held by it with a view to frustrate its disclosure. So if I am working for a public authority and I pretend that a record isn’t held in order to prevent an FOI punter from receiving it, I have committed an offence. If the organisation conspires in this, the organisation can itself be prosecuted by the Commissioner.

Back to my request to the ICO. They replied (once again, remarkably close to the 20 day deadline), and told me two interesting things. First, in answer to my question about whether the policy was held in a final approved form: “The policy was not held in final approved form“. Second, any recorded information about whether any data held constituted the requested information, or whether the ICO should in fact responded that the information was not held: “We do not hold recorded information. As you will be aware the Freedom of Information Act only covers recorded information held by a public authority. However, it may help you to know that there was a verbal discussion in regard to the response to this Freedom of Information request.” So, there was a verbal discussion that people plainly remember, and the ICO thinks it might help me to know this, without even a squeak about what the discussion was about. Thanks, Wilmslow, consider me unenlightened.

I believe that the ICO’s response to McKee’s request is untrue. The correct answer to her request is ‘no information held’, with advice and assistance that the data was in draft. Section 22 applies where the requested information exists but the organisation intends to publish it unchanged in the future; the ICO’s policy wasn’t complete. Look at what McKee asked for all those months ago: she asked for “your Policy designed to show compliance with Schedule 1, Part 4 of DPA 2018“. An incomplete, unapproved policy plainly does not answer the request, and the ICO should have confirmed that. The use of the exemption was a dishonest dodge to avoid admitting the truth.

If the ICO had a policy and pretended that they did not, under Section 77 it would have been a criminal offence for them to conceal its existence once it had been requested. As it happens, the ICO did the opposite – pretending that the information existed and refusing to give it out because it would be published in the future, rather than admitting that several months after the DPA was passed, the policy was not complete. Whoever decided that this was the right approach should think long and hard about a transparency regulator taking such a cynical attitude to legislation they are supposed to uphold and protect.

While QE2 tries to grab the headlines, demanding that FOI be extended to cover new organisations, her own house is far from being in order. The lack of FOI enforcement against recalcitrant and secretive government departments is an ongoing stain on the ICO’s reputation, while the lazy cynicism and lack of frankness over the office’s own activities suggests that the ICO can talk the talk, but walking the walk is beyond them. Regular readers of this blog are probably inured to my lack of faith in House Wycliffe, but for all Denham’s chasing of headlines, day to day experience of how the ICO carries out the most mundane of its functions suggests carelessness and disarray. Rather than trumpeting the press releases about extending FOI to charities and commercial bodies, more people should ask whether the ICO is capable of doing even those tasks it already has.

Regulating the FOIA into obscurity?

This is a guest post from the redoubtable John Slater, whose tireless efforts to hold DWP to account are a lesson in how FOI should be used. John has had real success in wrestling information out of a stubborn and secretive system, but the post describes the hurdles in the way of the applicant, and the shameful way in which the ICO makes things worse. It’s not a quick read but there’s a lot to say. I think anyone with an interest in how the benefits system operates, or how healthy the FOI system is at the moment should give it the time it deserves. I’m very grateful to John for writing it and letting me host it.

I suspect that most people reading this have experience of submitting a request for information (“RFI”) under the FOIA and all the frustrations that can come with it. Some people may have complained to the office of the Information Commissioner (“ICO”) while others may have just given up when their RFI was refused. I suspect that a smaller number of people, who had the time, appealed ICO decisions to the First-Tier and Upper Tribunals.

Via my involvement with the FOIA I have been dealing with the ICO for approximately 6 years. My interaction has ranged from normal FOIA complaints through to appeals to the First-Tier and Upper Tribunals.

Setting aside the minor issues one typically experiences with any large organization I have to say that my experience of dealing with the ICO has been very positive. Even when a decision notice (“DN”) went against me I could understand why and how that decision was reached. In respect of appeals to the First-Tier and Upper Tribunals I have nothing but praise for the people involved, even when I was appealing an ICO decision.

However, approximately 18 months ago things started to change for the worse. The time taken to respond to complaints seems to be inexorably increasing and the quality of the case work is deteriorating. I’ll use 3 of my current complaints to illustrate the problems that I and others are experiencing on a regular basis.

Case 1 – Universal Credit Programme Board Information Packs

In July 2017 I asked the DWP for the 3 most recent packs of information that were given to the Universal Credit (“UC”) Programme Board members at each monthly meeting. Given how controversial UC is and the history of the DWP being less than honest about it, this seemed to be a good route to try to find out what the senior people responsible for UC actually know and what they are doing about it.

For those not familiar with programme management terminology the programme board consists of senior people who are accountable and responsible for the UC programme, defining the direction of the programme and establishing frameworks to achieve its objectives. So apart from Neil Couling (senior responsible owner) and the secretary of state they are about as senior as it gets. The membership of the programme board can be found here:

https://www.whatdotheyknow.com/request/419990/response/1090823/attach/html/2/3044%20IR%20516%20IR%20604%20reply.pdf.html

Unsurprisingly the DWP refused my RFI on 16 August 2017 citing S.36. However it explained that it needed an extension to carry out the public interest test (“PIT”). On 14 September 2017 the DWP did exactly the same thing. This is a tactic that the DWP uses regularly and often issues monthly PIT extensions until the ICO becomes involved.

I complained to the ICO on 14 September 2017. On 22 November a DN was issued giving the DWP 35 calendar days to issue its response. On 3 January 2018 the DWP finally confirmed that it was engaging S.36 and that the public interest did not favour disclosure (I’ve yet to see a public interest test from the DWP that does favour disclosure). I submitted a revised complaint to the ICO on 9 January 2018 challenging S.36 and the public interest decision.

Despite the 5 month delay by the DWP the ICO bizarrely told me that I still had to exhaust the DWP internal review procedure before my complaint could be investigated. I had submitted 4 internal review requests (“IRR”) during the 5 months that the DWP treated the FOIA with such contempt. I know from previous experience that the DWP would use the same PIT ‘trick’ to delay answering my IRR. I explained this to the ICO and asserted that it has the authority to proceed without me having to submit another IRR. On 30 January the ICO accepted my complaint. I know about this from experience but I assume most people would have followed the ICO instruction and been stuck in another loop of 5 months until the DWP was told to issue its response to the IRR.

On 26 April my case was assigned to a case officer, just 3 months short of a year since I submitted my request to the DWP. Despite the DWP clearly citing S.36 the ICO allowed the DWP to get away with numerous delaying tactics and nothing happened for many months. Despite chasing the ICO on a number of occasions there appeared to be no progress. My patience ran out in October 2018 and I complained to the ICO about this and two other cases. On the face of it this appeared to have got things moving.

However, on 18 October 2018 I was told by the ICO that an information notice had been served on the DWP to obtain copies of the information I had requested. The DWP has 30 days to respond to these notices.

Whilst I’m not surprised by this (in fact I even suggested this was the case in my complaint) I struggle to understand how any organisation can investigate a complaint for almost 6 months without having a copy of the requested information. I can only hope that the DN I have been seeking for so long will appear at some point in 2018!

The delay has been so long that I have actually submitted another request for more current programme board packs. At the time of writing the DWP hasn’t provided a response within 20 days so that’s another complaint that I need to send to the ICO!

Case 2 – Aggregation of various RFIs

Between 4 February and 23 April 2018 the DWP aggregated 9 of my requests for information claiming that they were for the “same or similar” information. Well, what it actually said was:

We consider each of the seven requests to be of a similar nature as they all relate to either decision making or performance delivery of disability assessments on behalf of the Department for Work and Pensions.  In particular, all of the requests would be allocated to the same team for response as it falls within their specialised area. 

Under Section 12 of the FOI Act the Department is not therefore obliged to comply with your request and we will not be processing it further.

This seems to suggest that the DWP believes the requested information is the same or similar because they relate to activities it carries out and the teams that do them. This is a crude attempt to rely on the discredited concept of ‘overarching themes’ that was attempted in Benson v IC and the Governing Body of Buckinghamshire New University (EA20110016).  At [29] the Tribunal stated:

Whilst the Tribunal understood the Commissioner’s analysis the Tribunal felt that it was not compelling and relied on concepts that were not actually within the legislation – e.g. ‘overarching theme’. The Tribunal felt that any consequent uncertainty should, on balance, be resolved in the Appellant’s favour.

On 30 March I submitted a complaint to the ICO. My complaint involves 9 requests and deals with an important area of the FOIA, where there is very little precedent. A reasonable person might conclude that the ICO would be keen to act swiftly. On 27 April 2018 my complaint was assigned to a case officer so things were looking good. It is now coming towards the end of October and I have not had a single piece of correspondence from the ICO.

The requests that have been aggregated cover management information about how the DWP runs large controversial contracts that assess the eligibility for employment support allowance and personal independence payment (“PIP”). A previous RFI uncovered numerous problems with the quality of medical reports being produced for PIP assessments. This might explain why the DWP is so keen not to let me have the current information but not why there has been no progress by the ICO.

Case 3 – Datasets & Type of Data Held for Various Benefits About Claimants

On 26 February 2018 I asked the DWP to disclose the datasets and type of data it holds about various social security benefits. I am not asking for the actual data just the type of data and the “groups” or “sets” of data that it holds.

On 17 April 2018 the DWP refused my request citing S.31 (it eventually confirmed it meant section 31(1)(a))  and  S.24. After a further IRR the DWP reconfirmed its position and I complained to the ICO on 15 July. Some 3 months later on 11 October I was finally told that my case had been assigned to a case officer. Does this now mean I wait for a further 6 months before anything actually happens?

Conclusion

I know the ICO is very busy, partially due to the new Data Protection legislation, but the problems that I and others are experiencing can’t just be explained by “being busy”. Based on my previous experience of dealing with them I also don’t believe it is the fault of the case officers. These problems are due to serious organisational failings within the ICO. There doesn’t seem to be the type of business processes / workflow that one would expect to see in an organisation of this size. The line management oversight of case officers appears to be absent. Based on my own experience it seems to be that the line managers focus solely on protecting case officers while actually making matters worse for them as their workloads probably grow faster than they can cope with.

The ICO should have a small set of metrics about how it is dealing with cases. Surely line managers should be looking at cases where nothing has actually happened for 6 months and do something about it? The idea of management by exception has been around for a long time and yet I’m left with the impression that there are no exceptions set within the ICO and senior management have no impartial way of knowing what is actually going on at the case level.

People might wonder why this matters and that in these times of constrained budgets we should expect cases to take longer. I can’t accept this as one of the key drivers for the FOIA is that we get a chance to hold public authorities to account for their actions. For that to happen we need access to information while it is still relatively current.

It is generally known that there are certain large government departments that have very poor history in respect of FOIA. If someone requests information that these departments suspect will be embarrassing they will deliberately play the system to delay disclosure. From personal experience it’s all far too easy to do:

  1. Ignore the request completely until the ICO tells the department to respond (3+ months).
  2. Use the public interest test with impunity to introduce a 5 to 6 month delay before the requester can complain to the ICO about the exemption cited.
  3. 3 months before a case officer is assigned.
  4. At least 3 to 6 months before a DN is issued.

Total possible delay = 14 to 18 months.

The department can then appeal the DN to the First-Tier Tribunal (“FTT”), even if there is little chance of success. I’ve had 2 cases recently that have been appealed and then withdrawn just before the FTT hearing was due to take place. This added another 6 month delay let alone the cost to the public purse. If the DWP had actually gone through with the appeals and lost then that delay would probably be closer to 9 to 12 months.

This means that “playing the system” allows disreputable government departments to delay disclosure of embarrassing information by at least 2 years. Any media interest in the information can then be met with the claim that it is now ‘historical’ and things are better now.

A good example of this is the Project Assessment Review Reports (“PARs”) for the Universal Credit programme. I asked the DWP for these in April 2016 (see URL below):

https://www.whatdotheyknow.com/request/universal_credit_programme_proje#comment-82746

Using the delaying tactics described above and making the ICO issue an information notice to compel the DWP to release the PARs to them, they weren’t disclosed until March 2018. That’s a 2 year delay.

The ICO needs to sort out the internal delays that these government departments seem to be relying on. They also need to make sure there are meaningful consequences for public authorities that “play the system”. Writing strongly worded DNs telling public authorities off for abusing the system is meaningless. The ICO was highly critical of the DWP in its DN for the PARs case. A link to the DN is given below and the criticisms start at [62].

https://ico.org.uk/media/action-weve-taken/decision-notices/2017/2014762/fs50640285.pdf

The criticism has had absolutely no impact on the DWP.  It still regularly doesn’t reply in time and still produces “boilerplate” responses that have little bearing on the case in question.

As a result of the new GDPR and Facebook the Information Commissioner regularly seems to be in the media and was recently named as the most influential person in data-driven business in the updated DataIQ 100 list. I hear talk of the Commissioner being able to issue huge fines for data breaches and serving enforcement notices on organisations that are not complying with the FOIA.

The original white paper “your right to know” stated at [1.1]:

Unnecessary secrecy in Government leads to arrogance in government and defective decision-making. The perception of excess secrecy has become a corrosive influence in the decline of public confidence. Moreover, the climate of public opinion has changed; people expect much greater openness and accountability from government than they used to.”

If public authorities continue to be allowed to easily introduce delays of 2 years before disclosure then the regulator of the FOIA is failing in her role.  Before the FOIA we only had the thirty-year rule (now moving to the twenty-year rule) controlling when information was released to the public.

I suggest that we are rapidly approaching the situation where by default we have the “two-year rule” for information government departments do not want released. Unless the Commissioner does something about it that will slowly increase to the “three-year rule” and then the “four-year rule”. From my perspective its time the Commissioner stopped boasting about all the powers she has and started using them.

Live and Let Dai

To say that anything connected with GDPR is the worst example of its kind is a foolhardy business. I’ve read so many terrible articles, LinkedIn posts and Tweets about GDPR, to single any one of them out and say ‘THIS ONE IS THE WORST’ seems pointless. Most of them are bad. However, after watching 33 minutes of waffle, padding and gleefully misinformed bullshit, I am reckless enough to say that the intellectual property lawyer Dai Davis’ talk here is the worst presentation or talk I have seen about the GDPR in any format.

Admittedly, the trainer in me hated it because of the incompetence – Davis has to keep going back to the podium to change slides because he hasn’t brought a remote, and he pads the talk out with protracted questions to the audience that don’t add anything to what he is saying. When someone intelligent-sounding in the audience takes him on by asking a proper question, he runs a mile.

More seriously, a good chunk of the talk is taken up with an attempt to create a formula for how much you should spend on data protection compliance based on the likelihood of being fined. It’s an eye-catching and controversial thing to throw out in a conference, but I don’t believe even Davis knows what point he’s making. Is he really saying that a every organisation should spend a meaningless, averaged-out €2000 to comply with GDPR, or is that just a flourish? Every organisation is different to another, and will have radically different priorities and appetites for risk, so trying to create a standardised methodology is so random and unhelpful, I don’t think it’s a serious point.  Given the number of basic mistakes and baseless assertions he makes in such a short time, however, the only thing I can add to his calculations is that however much you spend on GDPR, you should probably not spend it on advice from him.

I may not have got them all, but here is as full a collection of all the blunders as I could manage:

  • Davis cannot remember how many deputies the Commissioner has, but he knows that it’s between 11 and 13. There are 3 deputies (James Dipple-Johnstone, Paul Arnold and Steve Wood); there have never been more than 3.
  • Davis consistently gets the name of the ICO wrong – it’s almost always the ‘Information Commission Office’, although he varies it at least once with ‘Information Commission Data Protection Officer’ (he wasn’t talking about their DPO). To be charitable, it might be because he’s talking quickly, but the errors are relentless. He clearly thinks that Elizabeth Denham’s job title is ‘ICO’. because he calls her this repeatedly, and talks about what he would do if he was “the ICO“.
  • He asserts that the GDPR is not a ‘step change’ from the old legislation solely because it has lots of words, even though many of those words are very similar to words in the same order in the old version
  • He notes that there has not been a GDPR fine yet. Davis was speaking on May 30th, two days after the first 72 hours to *report* a relevant breach would have elapsed.
  • He asserts several times that in theory “every single breach” has to be reported to the ICO. This is completely false. There is a specific definition of a breach in the GDPR and incidents that do not meet a certain threshold of risk do not have to be reported.
  • He says that telecoms companies had to report breaches to the ICO since 2012. Communications providers have had this duty since 2011, not just telecoms companies.
  • Davis claims that public sector bodies self-report breaches to the ICO because they have no idea about how to take a commercial risk. There is the problem that public sector bodies are not commercial organisations by and large, so that argument makes no sense, but it’s also factually incorrect. To take one example, NHS bodies (the example shouted out by an audience member) have been obliged by the operation of the Information Governance Toolkit to report breaches to the ICO since at least 1st June 2013 (I think it was actually earlier than this, but that’s the one given in a Toolkit document that Davis could have found with a single Google search if facts were something he had any curiosity about).
  • Davis claims that the ICO is not really responsible for prosecutions for S55 offences, despite talking exclusively about prosecutions that the ICO carried out.
  • He includes the prosecutions in his calculations for the risk of being fined by the ICO, seemingly unaware that fines and prosecutions are two entirely distinct activities, with S55 prosecutions being against individuals rather than organisations. Throughout, Davis talks about the ICO enforcing on ‘people’, so I don’t know if he knows that the penalties were issued against data controllers.
  • He says that there were 18000 complaints in 2016 and the ICO has done nothing about nearly all of them. As someone who thinks the ICO is crap, even I have to acknowledge that most of these complaints were resolved informally and the absence of a fine does not mean that nothing happened. In quite a few cases, the complaint would not have been valid, and so no action would be appropriate.
  • He twice says that the maximum penalty for a breach under the DPA 1998 was £5,000,000; it was £500,000.
  • He quotes the head of the ICO’s ‘Breach Notification Division’, which does not exist.
  • He claims that the GDPR contains more loopholes that requires the ICO to hire criminal lawyers. The standard of evidence for a GDPR breach is balance of probabilities, and GDPR removes the requirement to prove damage or distress for a monetary penalty.
  • He says the ICO has 700 staff – they haven’t recruited these staff yet.
  • He tells a story of how he tells his hotel clients (who, if they exist, have my pity) that they cannot claim to be GDPR compliant because they use “mobile telephones” and allow their staff to send text messages. According to Davis, it is impossible to use mobile phones securely.

At the point where Davis says “smart lawyers like me“, my jaw did not drop, it fell off.

Leaving aside how garbled and smug Davis’ performance is, you might wish to charitable and take on his central thesis – that you probably won’t get a GDPR fine. He’s right. There have been relatively few penalties under Data Protection thus far and so the risk of getting one is relatively small. I cannot disagree with this banal point because I have made it myself any times. However, I can’t tell if his conclusion is simply that nobody should bother complying or whether there would have been a ‘however, you should comply because…’ moment, because there isn’t a conclusion. Presumably because he has run out of time, Davis just stops. So what, Dai? What’s your point? What should the audience do with this information? Should they just ignore GDPR?  There’s definitely a sense of this when he says that 10 years from now, the owner of a B&B will not know what GDPR is.

If Davis had the guts or the discipline to get to a conclusion that GDPR doesn’t matter, that would have been something. His contempt for detail would still be an impediment, but ‘Ignore GDPR’ is an assertion worth tackling. I could counter by arguing that the threat of a fine isn’t a good reason to comply, but respecting human dignity and avoiding harm to real people though inaccuracy, intrusion and insecurity is, but Davis never stops circling the airport, so I don’t even know if that’s what he’s saying.

If his contention that organisations don’t have the ability to measure risk effectively and need to get GDPR in perspective, that’s actually a good point, but he makes it so incompetently that again I’m not motivated to take him on. I have grudging sympathy for the idea that reputational damage is an overhyped risk (again, it’s not a point he makes clearly), but I know that many in the Data Protection world would passionately disagree, and I suspect that they could use Facebook’s current woes as evidence that public perception over data misuse isn’t something that boardrooms can ignore.

In the end, I think Davis is a clever man pontificating about a subject he neither cares for or understands, but the danger is that people will watch the talk and be contaminated by it. You could argue that I am making it worse by drawing attention to it solely so I can take the piss. All I can say is, the talk is out there. People will see it. As this is the case, if you find his argument (such as it is) attractive, it’s worth pointing out how sloppy and ill-informed his thinking is. It’s worth asking if this is the ‘Ignore GDPR’ guy, why would you listen to him?

Cop out

On May 3rd 2018, Elizabeth Denham appeared on Channel 4 News as part of her long running commitment to generating headlines. Denham’s track record on the programme is not great – it was on the same programme in March that she adopted the interesting tactic (uniquely, as far as I can see) of informing an organisation in public and in advance that she planned to apply for a warrant to raid them, losing what might be a useful element of surprise in order to look tough in front of Jon Snow.

In the more recent interview, the Commissioner claimed that she had the power to fine directors and had done so. I made an FOI request about this, and the ICO admitted that “we do not have the power to directly fine directors“, directly contradicting what Denham said. You can tell me that ICO has the power to go after directors in limited circumstances that can result in a court issuing a fine and that must be what she meant (ICO did) but that’s not good enough. The DP regulator went on the telly and claimed to have a power she doesn’t have – it’s surely part of Denham’s job to increase understanding of Data Protection, not to muddy the waters.

In the same interview, Denham cheerily announced that she saw herself as a Sheriff of the internet. Arguably, she should be a Mountie but let’s leave that to one side. I assumed that the statement was a throwaway, not a serious statement of how Denham sees herself and her office. I was wrong. There’s a pattern. In a fawning profile by the Observer’s Carole Cadwalladr a few weeks ago, the Commissioner delivered a soundbite that I suspect is intended to epitomise the Denham Era: “Data crimes are real crimes“. And in the recently leaked DCMS Committee report into Fake News, she was at it again:

For the public, we need to be able to understand why an individual sees a certain ad. Why does an individual see a message in their newsfeed that somebody else does not see? We are really the data cops here. We are doing a data audit to be able to understand and to pull back the curtain on the advertising model around political campaigning and election

I think the misleading impression being created here could attract the label ‘fake news’ just as much as any of the internet nonsense Denham and her fanbase are supposedly against. Data crimes are usually not real crimes, and in most cases, the ICO are not the cops. The GDPR doesn’t make anything a criminal offence, and the offences under the Data Protection Act 2018, like those in its predecessor the 1998 Act, are specific. It’s a criminal offence to take, procure or sell personal data without the permission of the data controller; it’s an offence to re-identify depersonalised data (in circumstances so tightly defined I doubt there will be a successful prosecution), and it can be an offence to oblige someone to make a subject access request. Admittedly, the DPA 2018 is stricter in this area – offences under the DPA 1998 were not recordable so you wouldn’t get a criminal record if you committed them, a position that is sensibly reversed in the new version.

However, in some circumstances, the DPA 2018 is less oriented towards offences than the  DPA 1998. A breach of an Enforcement or Information Notice is no longer subject to prosecution, being punishable by a penalty instead. That might result in stricter punishments, but that depends on Wilmslow showing a willingness to use the powers, and in any case, it’s not a criminal sanction. The much-vaunted criminal prosecution of SCL by the Commissioner over David Carroll’s subject access request is doomed in my opinion, but if it goes ahead, it will almost certainly be the last prosecution for a breach of a notice. None of the DP offences are punishable with prison, and for all Denham’s bluster about being a data cop, she never publicly applies the pressure for custodial sentences. For all his faults, her predecessor Christopher Graham never missed an opportunity to do so.

If Facebook willingly shared its customers personal data with Cambridge Analytica, it would not be a criminal offence. If they reused their customers’ data and sold it to list brokers, it would not be a criminal offence. As drafted, the ‘victim’ of most data protection offences would be the data controller, not the person whose data is misappropriated, sold or misused. Denham wants to conjure up images of cops and robbers, but she’s misleading the public. Who knows, maybe she doesn’t want people to realise that the only sanction for the majority of data transgressions are monetary penalty that she has the power to approve. Maybe she means ‘data crimes should be real crimes‘, but if that’s the case, that what she should say instead of giving the wrong impression.

There’s another problem. By setting herself up as the Internet Sheriff, Denham is creating expectations I don’t believe she’s prepared to meet. In all her public appearances, the Commissioner is clearly trying to mark out the internet and new technology as her manor. Supporters like Cadwalladr are only too happy to play along. The Observer piece contains a brief but devastating verdict on thirty or so years of ICO work and four previous Commissioners: “a somewhat dusty regulator dealing in a niche topic“. I’m the last person to defend the ICO, but this writes off Wilmslow’s endeavours on phone hacking, union blacklisting, the lost HMRC data disks and many DP and PECR fines which even I can’t deny have changed behaviour for the better in many sectors. I can’t say that Denham endorses this trashing of her predecessors’ efforts, but she hasn’t repudiated it either. What must her staff think of it?

Strip away the recent headlines for prosecutions and £500,000 fines that haven’t actually happened yet, and Denham’s record is hardly the Data Protection equivalent of Wyatt Earp taking on the Clantons. When dealing with the misuse of 1.6 million people’s data by the Royal Free Hospital and the AI company owned by Google (exactly the kind of tech territory we’re supposed to believe she wants to police), Denham’s ICO asked the Royal Free to sign an undertaking. There is no automatic sanction if they go back on it. Faced with multiple instances of charities profiling potential donors in secret (not a million miles away from the kind of surreptitious data gathering that attracts her current ire), Denham’s response was reportedly to cut the originally proposed fines, such that Oxfam was fined just £6000. Late in 2017, Sheriff Denham issued an enforcement notice against the Ministry of Justice over shameful and long-running subject access backlogs that doubtlessly affected many people in desperate legal circumstances. She gave them eight months to comply and sneaked the notice out on the last working day before Christmas without a press release.

You can tell me that the ICO has consistently issued monetary penalties on Denham’s watch but so did Graham, though the double whammy of £400,000 CMPs on both TalkTalk and Carphone Warehouse weigh against my argument to some extent. But beyond those, Denham has done nothing revolutionary or interesting in enforcement. There has been no action on accuracy or retention, and little on the vital first principle beyond the charity cases that were obviously started under Graham.

Outwardly, Denham seems poised and plausible. Fate has dealt her the biggest data protection story in a decade and some overly sympathetic press coverage, so maybe she’s right to milk it and build up her part. There’s no question that she has a higher public profile than any of the Commissioners who have gone before her, and I know a lot of people in the DP world who think that this is automatically a good thing. I’m not convinced. I think ‘data crimes are real crimes’ could become as unhelpful a distraction as the pervasive ‘GDPR = consent’ myth, and nothing about the past two years convinces me that Denham really has what it takes to round up the internet’s outlaws. As always, I will delighted to be proved wrong; some eyecatching monster scalps is what I have spent years of blogging asking for, and it will make my job easier for the next few years. But unless she really pulls out the big guns, the Commissioner’s legacy may be less Gunfight at the IT Corral, and more Denham’s Last Stand.