Category Archives: ethics

care.data, children, datasharing, ethics, privacy

Leaving Facebook and flaws in Face Recognition

This Facebook ad was the final straw for me this week.

I’m finally leaving.

When I saw Facebook’s disingenuous appropriation of new data law as-a-good-thing I decided time’s up. While Zuckerberg talks about giving users more control, what they are doing is steering users away from better privacy and putting users outside the reach of new protections rather than stepping up to meet its obligations.

After eleven years, I’m done. I’ve used Facebook to run a business.  I’ve used it to keep in touch with real-life family and friends. I’ve had more positive than negative experiences on the site. But I’ve packed in my personal account.

I hadn’t actively used it since 2015. My final post that year was about Acxiom’s data broker agreement with Facebook. It has taken 3 hours to download  any remaining data, to review and remove others’ tags, posts and shared content linking me. I had already deactivated 18 apps, and have now used each individual ID that the Facebook-App link provided, to make Subject Access requests (SAR) and object to processing. Some were easy. Some weren’t.

Pinterest and Hootsuite were painful circular loops of online ‘support’ that didn’t offer any easy way to contact them.  But to their credit Hootsuite Twitter message support was ultra fast and suggested an email to hootsuite-dpa [at] hootsuite.com. Amazon required a log in to the Amazon account. Apple’s Aperture goes into a huge general page impossible to find any easy link to contact.  Ditto Networked Blogs.

Another app that has no name offered a link direct to a pre-filled form with no contact details and no option for free text you can send only the message please delete any data you hold about me — not make a SAR.

Another has a policy but no Data Controller listed. Who is http://a.pgtb.me/privacy ? Ideas welcome.

What about our personal data rights?

The Facebook ad says, you will be able to access, download or delete your data at any time. Not according to the definition of personal data we won’t.  And Facebook knows it. As Facebook’s new terms and condition says, some things that you do on Facebook aren’t stored in your account. For example, a friend may have messages from you after deletion. They don’t even mention data inferred. This information remains after you delete your account. It’s not ‘your’ data because it belongs to the poster, it seems according to Facebook. But it’s ‘your’ data because the data are about or related to you according to data protection law.

Rights are not about ownership.

That’s what Facebook appears to want to fail to understand. Or perhaps wants the reader to fail to understand. Subject Access requests should reveal this kind of data, and we all have a right to know what the Facebook user interface limits-by-design. But Facebook still keeps this hidden, while saying we have control.

Meanwhile, what is it doing?  Facebook appears to be running scared and removing  recourse to better rights.

Facebook, GDPR and flaws in Face Recognition

They’ve also started running Face Recognition. With the new feature enabled, you’re notified if you appear in a photo even if not tagged.

How will we be notified if we’re not tagged? Presumably Facebook uses previously stored facial images that were tagged, and is matching them using an image library behind the scenes.

In the past I have been mildly annoyed when friends who should know me better, have posted photos of my children on Facebook.

Moments like children’s birthday parties can mean a photo posted of ten fun-filled faces in which ten parents are tagged. Until everyone knew I’d rather they didn’t, I was often  tagged in photos of my young children.  Or rather my children were tagged as me.

Depending on your settings, you’ll receive a notification when someone tags a photo with your name.  Sure I can go and untag it, to change the audience that can see it, but cannot have control over it.

Facebook meanwhile pushes this back as if it is a flaw with the user and in a classic victim-blaming move suggests it’s your fault you don’t like it, not their failure to meet privacy-by-design, by saying,  If you don’t like something you’re tagged in, you can remove the tag or ask the person who tagged you to remove the post.

There is an illusion of control being given to the user, by companies and government at the moment. We must not let that illusion become the accepted norm.

Children whose parents are not on the site cannot get notifications. A parent may have no Facebook account.  (A child under 13 should no Facebook account, although Facebook has tried to grab those too.) The child with no account may never know, but Facebook is certainly processing, and might be building up a shadow profile about, the nameless child with face X anyway.

What happens next?

As GDPR requires a share of accountability for controller and processing responsibilities, what will it mean for posters who do so without consent of the people in photos? For Facebook it should mean they cannot process using biometric profiling, and its significant effects may be hidden or, especially for children, only appear in the future.

Does Facebook process across photos held on other platforms?

Since it was founded, Facebook has taken over several social media companies, the most familiar of which are Instagram in 2012 and WhatsApp in 2014. Facebook has also bought Oculus VR [VR headsets], Ascenta [drones], and ProtoGeo Oy [fitness trackers].

Bloomberg reported at the end of February that  a lawsuit alleging Facebook Inc. photo scanning technology flouts users’ privacy rights can proceed.

As TechCrunch summarised, when asked to clear a higher bar for privacy, Facebook has instead delved into design tricks to keep from losing our data.

Facebook needs to axe Face Recognition, or make it work in ways that are lawful, to face up to its responsibilities, and fast.

The Cambridge Analytica scandal has also brought personalised content targeting into the spotlight, but we are yet to see really constructive steps to row back to more straightfoward advertising, and away from todays’s highly invasive models of data collection and content micro-targeting designed to to grab your personalised attention.

Meanwhile policy makers and media are obsessed with screen time limits as a misplaced, over-simplified solution to complex problems, in young people using social media, which are more commonly likely to be exacerbating existing conditions and demonstrate correlations rather than cause.

Children are stuck in the middle.

Their rights to protection, privacy, reputation and participation must not become a political playground.

ethics, politics, research

When is a profile no longer personal data

This is bothering me in current and future data protection.

When is a profile no longer personal?

I’m thinking of a class, or even a year group of school children.

If you strip off enough identifiers or aggregate data so that individuals are no longer recognisable and could not be identified from other data — directly or indirectly — that is in your control or may come into your control, you no longer process personal data.

How far does Article 4(1) go to the boundary of what is identifiable on economic, cultural or social identity?

There is a growing number of research projects using public sector data (including education but often in conjunction and linkage with various others) in which personal data are used to profile and identify sets of characteristics, with a view to intervention.

Let’s take a case study.

Exclusions and absence, poverty, ethnicity, language, SEN and health, attainment indicators, their birth year and postcode areas are all profiled on individual level data in a data set of 100 London schools all to identify the characteristics of children more likely than others to drop out.

It’s a research project, with a view to shaping a NEET intervention program in early education (Not in Education, Employment or Training). There is no consent sought for using the education, health, probation, Police National Computer, and HMRC data like this, because it’s research, and enjoys an exemption.

Among the data collected BAME ethnicity and non-English language students in certain home postcodes are more prevalent. The names of pupils and DOB and their school address have been removed.

In what is in effect a training dataset, to teach the researchers, “what does a potential NEET look like?” pupils with characteristics like Mohammed Jones, are more likely than others to be prevalent.

It does not permit the identification of the data subject as himself, but the data knows exactly what a pupil like MJ looks like.

Armed with these profiles of what potential NEETs look like, researchers now work with the 100 London schools, to give the resulting knowledge, to help teachers identify their children at risk of potential drop out, or exclusion, or of becoming a NEET.

In one London school, MJ, is a perfect match for the profile. The teacher is relieved from any active judgement who should join the program, he’s a perfect match for what to look for.  He’s asked to attend a special intervention group, to identify and work on his risk factors.

The data are accurate. His profile does match. But would he have gone on to become NEET?

Is this research, or was it a targeted intervention?

Are the tests for research exemptions met?

Is this profiling and automated decision-making?

If the teacher is asked to “OK” the list, but will not in practice edit it, does that make it exempt from the profiling restriction for children?

The GDPR also sets out the rules (at Article 6(4)) on factors a controller must take into account to assess whether a new processing purpose is compatible with the purpose for which the data were initially collected.

But if the processing is done only after the identifiers are removed that could identify MJ, not just someone like him, does it apply?

In a world that talks about ever greater personalisation, we are in fact being treated less and less as an individual, but instead we’re constantly assessed by comparison, and probability, how we measure up against other profiles of other people built up from historical data.

Then it is used to predict what someone with a similar profile would do, and therefore by inference, what we the individual would do.

What is the difference in reality, of having given the researchers all the education, health, probation, Police National Computer, and HMRC — as they had it — and then giving them the identifying school datasets with pupils’ named data, and saying “match them up.”

I worry that we are at great risk, in risk prediction, of not using the word research, to mean what we think it means.

And our children are unprotected from bias and unexpected consequences as a result.

change, children, datasharing, empowerment, ethics, pupils

Failing a generation is not what post-Brexit Britain needs

Basically Britain needs Prof. Brian Cox shaping education policy:

“If it were up to me I would increase pay and conditions and levels of responsibility and respect significantly, because it is an investment that would pay itself back many times over in the decades to come.”

Don’t use children as ‘measurement probes’ to test schools

What effect does using school exam results to reform the school system have on children? And what effect does it have on society?

Last autumn Ofqual published a report and their study on consistency of exam marking and metrics.

The report concluded that half of pupils in English Literature, as an example, are not awarded the “correct” grade on a particular exam paper due to marking inconsistencies and the design of the tests.
Given the complexity and sensitivity of the data, Ofqual concluded, it is essential that the metrics stand up to scrutiny and that there is a very clear understanding behind the meaning and application of any quality of marking.  They wrote that, “there are dangers that information from metrics (particularly when related to grade boundaries) could be used out of context.”

Context and accuracy are fundamental to the value of and trust in these tests. And at the moment, trust is not high in the system behind it. There must also be trust in policy behind the system.

This summer two sets of UK school tests, will come under scrutiny. GCSEs and SATS. The goal posts are moving for children and schools across the country. And it’s bad for children and bad for Britain.

Grades A-G will be swapped for numbers 1 -9

GCSE sitting 15-16 year olds will see their exams shift to a numerical system, scoring from the highest Grade 9 to Grade 1, with the three top grades replacing the current A and A*. The alphabetical grading system will be fully phased out by 2019.

The plans intended that roughly the same proportion of students as have achieved a Grade C will be awarded a new Grade 4 and as Schools Week reported: “There will be two GCSE pass rates in school performance tables.”

One will measure grade 5s or above, and this will be called the ‘strong’ pass rate. And the other will measure grade 4s or above, and this will be the ‘standard’ pass rate.

Laura McInerney summed up, “in some senses, it’s not a bad idea as it will mean it is easier to see if the measures are comparable. We can check if the ‘standard’ rate is better or worse over the next few years. (This is particularly good for the DfE who have been told off by the government watchdog for fiddling about with data so much that no one can tell if anything has worked anymore).”

There’s plenty of confusion in parents, how the numerical grading system will work. The confusion you can gauge in playground conversations, is also reflected nationally in a more measurable way.

Market research in a range of audiences – including businesses, head teachers, universities, colleges, parents and pupils – found that just 31 per cent of secondary school pupils and 30 per cent of parents were clear on the new numerical grading system.

So that’s a change in the GCSE grading structure. But why? If more differentiators are needed, why not add one or two more letters and shift grade boundaries? A policy need for these changes is unclear.

Machine marking is training on ten year olds

I wonder if any of the shift to numerical marking, is due in any part to a desire to move GCSEs in future to machine marking?

This year, ten and eleven year olds, children in their last year of primary school, will have their SATs tests computer marked.

That’s everything in maths and English. Not multiple choice papers or one word answers, but full written responses. If their f, b or g doesn’t look like the correct  letter in the correct place in the sentence, then it gains no marks.

Parents are concerned about children whose handwriting is awful, but their knowledge is not. How well can they hope to be assessed? If exams are increasingly machine marked out of sight, many sent to India, where is our oversight of the marking process and accuracy?

The concerns I’ve heard simply among local parents and staff, seem reflected in national discussions and the assessor, Oftsed. TES has reported Ofsted’s most senior officials as saying that the inspectorate is just as reluctant to use this year’s writing assessments as it was in 2016. Teachers and parents locally are united in feeling it is not accurate, not fair, and not right.

The content is also to be tougher.

How will we know what is being accurately measured and the accuracy of the metrics with content changes at the same time? How will we know if children didn’t make the mark, or if the marks were simply not awarded?

The accountability of the process is less than transparent to pupils and parents. We have little opportunity for Ofqual’s recommended scrutiny of these metrics, or the data behind the system on our kids.

Causation, correlation and why we should care

The real risk is that no one will be able to tell if there is an error, where it stems from, and where there is a reason if pass rates should be markedly different from what was expected.

After the wide range of changes across pupil attainment, exam content, school progress scores, and their interaction and dependencies, can they all fit together and be comparable with the past at all?

If the SATS are making lots of mistakes simply due to being bad at reading ten year’ old’s handwriting, how will we know?

Or if GCSE scores are lower, will we be able to see if it is because they have genuinely differentiated the results in a wider spread, and stretched out the fail, pass and top passes more strictly than before?

What is likely, is that this year’s set of children who were expecting As and A star at GCSE but fail to be the one of the two children nationally who get the new grade 9, will be disappointed to feel they are not, after all, as great as they thought they were.

And next year, if you can’t be the one or two to get the top mark, will the best simply stop stretching themselves and rest a bit easier, because, whatever, you won’t get that straight grade As anyway?

Even if children would not change behaviours were they to know, the target range scoring sent by third party data processors to schools, discourages teachers from stretching those at the top.

Politicians look for positive progress, but policies are changing that will increase the number of schools deemed to have failed. Why?

Our children’s results are being used to reform the school system.

Coasting and failing schools can be compelled to become academies.

Government policy on this forced academisation was rejected by popular revolt. It appears that the government is determined that schools *will* become academies with the same fervour that they *will* re-introduce grammar schools. Both are unevidenced and unwanted. But there is a workaround.  Create evidence. Make the successful scores harder to achieve, and more will be seen to fail.

A total of 282 secondary schools in England were deemed to be failing by the government this January, as they “have not met a new set of national standards”.

It is expected that even more will attain ‘less’ this summer. Tim Leunig, Chief Analyst & Chief Scientific Adviser Department for Education, made a personal guess at two reaching the top mark.

The context of this GCSE ‘failure’ is the changes in how schools are measured. Children’s progress over 8 subjects, or “P8” is being used as an accountability measure of overall school quality.

But it’s really just: “a school’s average Attainment 8 score adjusted for pupils’ Key Stage 2 attainment.” [Dave Thomson, Education Datalab]

Work done by FFT Education Datalab showed that contextualising P8 scores can lead to large changes for some schools.  (Read more here and here). You cannot meaningfully compare schools with different types of intake, but it appears that the government is determined to do so. Starting ever younger if new plans go ahead.

Data is being reshaped to tell stories to fit to policy.

Shaping children’s future

What this reshaping doesn’t factor in at all, is the labelling of a generation or more, with personal failure, from age ten and up.

All this tinkering with the data, isn’t just data.

It’s tinkering badly with our kids sense of self, their sense of achievement, aspiration, and with that; the country’s future.

Education reform has become the aim, and it has replaced the aims of education.

Post-Brexit Britain doesn’t need policy that delivers ideology. We don’t need “to use children as ‘measurement probes’ to test schools.”

Just as we shouldn’t use children’s educational path to test their net worth or cost to the economy. Or predict it in future.

Children’s education and human value cannot be measured in data.

care.data, change management, datasharing, engagement, ethics, HSCIC, human rights, NHS England, politics, privacy

A vanquished ghost returns as details of distress required in NHS opt out

It seems the ugly ghosts of care.data past were alive and well at NHS Digital this Christmas.

Old style thinking, the top-down patriarchal ‘no one who uses a public service should be allowed to opt out of sharing their records. Nor can people rely on their record being anonymised,‘ that you thought was vanquished, has returned with a vengeance.

The Secretary of State for Health, Jeremy Hunt, has reportedly  done a U-turn on opt out of the transfer of our medical records to third parties without consent.

That backtracks on what he said in Parliament on January 25th, 2014 on opt out of anonymous data transfers, despite the right to object in the NHS constitution [1].

So what’s the solution? If the new opt out methods aren’t working, then back to the old ones and making Section 10 requests? But it seems the Information Centre isn’t keen on making that work either.

All the data the HSCIC holds is sensitive and as such, its release risks patients’ significant harm or distress [2] so it shouldn’t be difficult to tell them to cease and desist, when it comes to data about you.

But how is NHS Digital responding to people who make the effort to write directly?

Someone who “got a very unhelpful reply” is being made to jump through hoops.

If anyone asks that their hospital data should not be used in any format and passed to third parties, that’s surely for them to decide.

Let’s take the case study of a woman who spoke to me during the whole care.data debacle who had been let down by the records system after rape. Her NHS records subsequently about her mental health care were inaccurate, and had led to her being denied the benefit of private health insurance at a new job.

Would she have to detail why selling her medical records would cause her distress? What level of detail is fair and who decides? The whole point is, you want to keep info confidential.

Should you have to state what you fear? “I have future distress, what you might do to me?” Once you lose control of data, it’s gone. Based on past planning secrecy and ideas for the future, like mashing up health data with retail loyalty cards as suggested at Strata in November 2013 [from 16:00] [2] no wonder people are sceptical. 

Given the long list of commercial companies,  charities, think tanks and others that passing out our sensitive data puts at risk and given the Information Centre’s past record, HSCIC might be grateful they have only opt out requests to deal with, and not millions of medical ethics court summonses. So far.

HSCIC / NHS Digital has extracted our identifiable records and has given them away, including for commercial product use, and continues give them away, without informing us. We’ve accepted Ministers’ statements and that a solution would be found. Two years on, patience wears thin.

“Without that external trust, we risk losing our public mandate and then cannot offer the vital insights that quality healthcare requires.”

— Sir Nick Partridge on publication of the audit report of 10% of 3,059 releases by the HSCIC between 2005-13

— Andy WIlliams said, “We want people to be certain their choices will be followed.”

Jeremy Hunt said everyone should be able to opt out of having their anonymised data used. David Cameron did too when the plan was  announced in 2012.

In 2014 the public was told there should be no more surprises. This latest response is not only a surprise but enormously disrespectful.

When you’re trying to rebuild trust, assuming that we accept that ‘is’ the aim, you can’t say one thing, and do another.  Perhaps the Department for Health doesn’t like the public answer to what the public wants from opt out, but that doesn’t make the DH view right.

Perhaps NHS Digital doesn’t want to deal with lots of individual opt out requests, that doesn’t make their refusal right.

Kingsley Manning recognised in July 2014, that the Information Centre “had made big mistakes over the last 10 years.” And there was “a once-in-a-generation chance to get it right.”

I didn’t think I’d have to move into the next one before they fix it.

The recent round of 2016 public feedback was the same as care.data 1.0. Respect nuanced opt outs and you will have all the identifiable public interest research data you want. Solutions must be better for other uses, opt out requests must be respected without distressing patients further in the process, and anonymous must mean  anonymous.

“Pseudonymised data requests that go through the DARS process so that a Data Sharing Framework Contract and Data Sharing Agreement are in place are considered to be compliant with the ICO code of practice“ – fine, but they are not anonymous. If DARS is still giving my family’s data to Experian, Harvey Walsh, and co, despite opt out, I’ll be furious.

The [Caldicott 2] Review Panel found “that commissioners do not need dispensation from confidentiality, human rights & data protection law.”

Neither do our politicians, their policies or ALBs.

[1] https://www.england.nhs.uk/ourwork/tsd/ig/ig-fair-process/further-info-gps/

“A patient can object to their confidential personal information from being disclosed out of the GP Practice and/or from being shared onwards by the HSCIC for non-direct care purposes (secondary purposes).”

[2] Minimum Mandatory Measures http://www.nationalarchives.gov.uk/documents/information-management/cross-govt-actions.pdf p7

AI, engagement, ethics, politics, privacy

Data for Policy: Ten takeaways from the conference

The knowledge and thinking on changing technology, the understanding of the computing experts and those familiar with data, must not stay within conference rooms and paywalls.

What role do data and policy play in a world of post-truth politics and press? How will young people become better informed for their future?

The data for policy conference this week, brought together some of the leading names in academia and a range of technologists, government representatives, people from the European Commission, and other global organisations, Think Tanks, civil society groups, companies, and individuals interested in data and statistics. Beyond the UK, speakers came from several other countries in Europe, from the US, South America and Australia.

The schedule was ambitious and wide-ranging in topics. There was brilliant thinking and applications of ideas. Theoretical and methodological discussions were outnumbered by the presentations that included practical applications or work in real-life scenarios using social science data, humanitarian data, urban planning, public population-wide administrative data from health, finance, documenting sexual violence and more. This was good.

We heard about lots of opportunities and applied projects where large datasets are being used to improve the world. But while I always come away from these events having learned something and encouraged to learn more about those I didn’t, I do wonder if the biggest challenges in data and policy aren’t still the simplest.

No matter how much information we have, we must use it wisely. I’ve captured ten takeaways of things I would like to see follow. This may not have been the forum for it.

Ten takeaways on Data-for-Policy

1. Getting beyond the Bubble

All this knowledge must reach beyond the bubble of academia, beyond a select few white-male-experts-in well off parts of the world, and get into the hands and heads of the many. Ways to do this must include reducing the cost or changing  pathways of academic print access. Event and conference fees are also a  barrier to many.

2. Context of accessibility and control

There is little discussion of the importance of context. The nuance of most of these subjects was too much for the length of the sessions but I didn’t hear any single session mention threats to data access and trust in data collection posed by surveillance or state censorship or restriction of access to data or information systems, or the editorial control of knowledge and news by Facebook and co. There was no discussion of the influence of machine manipulators, how bots change news or numbers and create fictitious followings.

Policy makers and public are influenced by the media, post-truth or not. Policy makers in the UK government recently wrote in response to challenge over a Statutory Instrument that if Mums-net wasn’t kicking up  a fuss then they believed the majority of the public were happy. How are policy makers being influenced by press or social media statistics without oversight or regulating for their accuracy?

Increasing data and technology literacy in policy makers, is going to go far beyond improving an understanding of data science.

3. Them and Us

I feel a growing disconnect between those ‘in the know’ and those in ‘the public’. Perhaps that is a side-effect of my own understanding growing about how policy is made, but it goes wider. Those who talked about ‘the public’ did so without mention that attendees are all part of that public. Big data, are often our data. We are the public.

Vast parts of the population feel left behind already by policy and government decision-making; divided by income, Internet access, housing, life opportunites, and the ability to realise our dreams.

How policy makers address this gulf in the short and long term both matter as a foundation for what data infrastructure we have access to, how well trusted it is, whose data are included and who is left out of access to the information or decision-making using it.

Researchers prevented from accessing data held by government departments, perhaps who fear it will be used to criticise rather than help improve policy of the day, may be limiting our true picture of some of this divide and its solutions.

Equally data that is used to implement top-down policy without public involvement, seems a shame to ignore public opinion. I would like to have asked, does GDS in its land survey work searching for free school sites include people surveys asking, do you want a free school in your area at all?

4. There is no neutral

Global trust in politics is in tatters. Trust in the media is as bad. Neither appear to be interested across the world in doing much to restore their integrity.

All the wisdom in the world could not convince a majority in the 23rd June referendum, that the UK should remain in the European Union. This unspoken context was perhaps an aside to most of the subjects of the conference which went beyond the UK,  but we cannot ignore that the UK is deep in political crisis in the world, and at home the Opposition seems to have gone into a tailspin.

What role do data and evidence have in post-truth politics?

It was clear in discussion, that if I mentioned technology and policy in a political context, eyes started to glaze over. Politics should not interfere with the public interest, but it does and cannot be ignored. In fact it is short term political terms and needs for long term vision that are perhaps most at-odds in making good data policy plans.

The concept of public good, is not uncomplicated. It is made more complex still if you factor in changes over time, and cannot ignore that Trump or Turkey are not fictitious backdrops considering who decides what the public good and policy priorities should be.

Researchers’ role in shaping public good is not only about being ethical in their own research, but having the vision to have safeguards in place for how the knowledge they create are used.

5. Ethics is our problem, but who has the solution?

While many speakers touched on the common themes of ethics and privacy in data collection and analytics, saying this is going to be one of our greatest challenges, few address how, and who is taking responsibility and accountability for making it happen in ways that are not left to big business and profit making decision-takers.

It appears from last year, that ethics played a more central role. A year later we now have two new ethical bodies in the UK, at the UK Statistics Authority and at the Turing Institute. How they will influence the wider ethics issues in data science remains to be seen.

Legislation and policy are not keeping pace with the purchasing power or potential of the big players, the Googles and Amazons and Microsofts, and a government that sees anything resulting in economic growth as good, is unlikely to be willing to regulate it.

How technology can be used and how it should be used still seems a far off debate that no one is willing to take on and hold policy makers to account for. Implementing legislation and policy underpinned with ethics must serve as a framework for giving individuals insight into how decisions about them were reached by machines, or the imbalance of power that commercial companies and state agencies have in our lives that comes from insights through privacy invasion.

6. Inclusion and bias

Clearly this is one event in a world of many events that address similar themes, but I do hope that the unequal balance in representation across the many diverse aspects of being human are being addressed elsewhere.  A wider audience must be inclusive. The talk by Jim Waldo on retaining data accuracy while preserving privacy was interesting as it showed how deidentified data can create bias in results if data is very different from the original. Gaps in data, especially using big population data which excludes certain communities, wasn’t something I heard discussed as much.

7.Commercial data sources

Government and governmental organisations appear to be starting to give significant weight to the use of commercial data and social media data sources. I guess any data seen as ‘freely available’ that can be mined seems valuable. I wonder however how this will shape the picture of our populations, with what measures of validity and  whether data are comparable and offer reproducability.

These questions will matter in shaping policy and what governments know about the public. And equally, they must consider those communities whether in the UK or in other countries, that are not represented in these datasets and how these bias decision-making.

8. Data is not a panacea for policy making

Overall my take away is the important role that data scientists have to remind policy makers that data is only information. Nothing new. We may be able to access different sources of data in different ways, and process it faster or differently from the past, but we cannot rely on data of itself to solve the universal problems of the human condition. Data must be of good integrity to be useful and valuable. Data must be only one part of the library of resources to be used in planning policy. The limitations of data must also be understood. The uncertainties and unknowns can be just as important as evidence.

9. Trust and transparency

Regulation and oversight matter but cannot be the only solutions offered to concerns about shaping what is possible to do versus what should be done. Talking about protecting trust is not enough. Organisations must become more trustworthy if trust levels are to change; through better privacy policies, through secure data portability and rights to revoke consent and delete outdated data.

10. Young people and involvement in their future

What inspired me most were the younger attendees presenting posters, especially the PhD student using data to provide evidence of sexual violence in El Salvador and their passion for improving lives.

We are still not talking about how to protect and promote privacy in the Internet of Things, where sensors on every street corner in Smart Cities gather data about where we have been, what we buy and who we are with. Even our children’s toys send data to others.

I’m still as determined to convince policy makers that young people’s data privacy and digital self-awareness must be prioritised.

Highlighting the policy and practice failings in the niche area of the National Pupil Database serves only to get ideas from others how  policy and practice could be better. 20 million school children’s records is not a bad place to start to make data practice better.

The questions that seem hardest to move forward are the simplest: how to involve everyone in what data and policy may bring for future and not leave out certain communities through carelessness.

If the public is not encouraged to understand how our own personal data are collected and used, how can we expect to grow great data scientists of the future? What uses of data put good uses at risk?

And we must make sure we don’t miss other things, while data takes up the time and focus of today’s policy makers and great minds alike.

cb-poster-for-web

AI, change management, empowerment, engagement, ethics, surveillance, Tech, transparency

Mum, are we there yet? Why should AI care.

Mike Loukides drew similarities between the current status of AI and children’s learning in an article I read this week.

The children I know are always curious to know where they are going, how long will it take, and how they will know when they get there. They ask others for guidance often.

Loukides wrote that if you look carefully at how humans learn, you see surprisingly little unsupervised learning.

If unsupervised learning is a prerequisite for general intelligence, but not the substance, what should we be looking for, he asked. It made me wonder is it also true that general intelligence is a prerequisite for unsupervised learning? And if so, what level of learning must AI achieve before it is capable of recursive self-improvement? What is AI being encouraged to look for as it learns, what is it learning as it looks?

What is AI looking for and how will it know when it gets there?

Loukides says he can imagine a toddler learning some rudiments of counting and addition on his or her own, but can’t imagine a child developing any sort of higher mathematics without a teacher.

I suggest a different starting point. I think children develop on their own, given a foundation. And if the foundation is accompanied by a purpose — to understand why they should learn to count, and why they should want to — and if they have the inspiration, incentive and  assets they’ll soon go off on their own, and outstrip your level of knowledge. That may or may not be with a teacher depending on what is available, cost, and how far they get compared with what they want to achieve.

It’s hard to learn something from scratch by yourself if you have no boundaries to set knowledge within and search for more, or to know when to stop when you have found it.

You’ve only to start an online course, get stuck, and try to find the solution through a search engine to know how hard it can be to find the answer if you don’t know what you’re looking for. You can’t type in search terms if you don’t know the right words to describe the problem.

I described this recently to a fellow codebar-goer, more experienced than me, and she pointed out something much better to me. Don’t search for the solution or describe what you’re trying to do, ask the search engine to find others with the same error message.

In effect she said, your search is wrong. Google knows the answer, but can’t tell you what you want to know, if you don’t ask it in the way it expects.

So what will AI expect from people and will it care if we dont know how to interrelate? How does AI best serve humankind and defined by whose point-of-view? Will AI serve only those who think most closely in AI style steps and language?  How will it serve those who don’t know how to talk about, or with it? AI won’t care if we don’t.

If as Loukides says, we humans are good at learning something and then applying that knowledge in a completely different area, it’s worth us thinking about how we are transferring our knowledge today to AI and how it learns from that. Not only what does AI learn in content and context, but what does it learn about learning?

His comparison of a toddler learning from parents — who in effect are ‘tagging’ objects through repetition of words while looking at images in a picture book — made me wonder how we will teach AI the benefit of learning? What incentive will it have to progress?

“the biggest project facing AI isn’t making the learning process faster and more efficient. It’s moving from machines that solve one problem very well (such as playing Go or generating imitation Rembrandts) to machines that are flexible and can solve many unrelated problems well, even problems they’ve never seen before.”

Is the skill to enable “transfer learning” what will matter most?

For AI to become truly useful, we need better as a global society to understand *where* it might best interface with our daily lives, and most importantly *why*.  And consider *who* is teaching and AI and who is being left out in the crowdsourcing of AI’s teaching.

Who is teaching AI what it needs to know?

The natural user interfaces for people to interact with today’s more common virtual assistants (Amazon’s Alexa, Apple’s Siri and Viv, Microsoft  and Cortana) are not just providing information to the user, but through its use, those systems are learning. I wonder what percentage of today’s  population is using these assistants, how representative are they, and what our AI assistants are being taught through their use? Tay was a swift lesson learned for Microsoft.

In helping shape what AI learns, what range of language it will use to develop its reference words and knowledge, society co-shapes what AI’s purpose will be —  and for AI providers to know what’s the point of selling it. So will this technology serve everyone?

Are providers counter-balancing what AI is currently learning from crowdsourcing, if the crowd is not representative of society?

So far we can only teach machines to make decisions based on what we already know, and what we can tell it to decide quickly against pre-known references using lots of data. Will your next image captcha, teach AI to separate the sloth from the pain-au-chocolat?

One of the task items for machine processing is better searches. Measurable goal driven tasks have boundaries, but who sets them? When does a computer know, if it’s found enough to make a decision. If the balance of material about the Holocaust on the web for example, were written by Holocaust deniers will AI know who is right? How will AI know what is trusted and by whose measure?

What will matter most is surely not going to be how to optimise knowledge transfer from human to AI — that is the baseline knowledge of supervised learning — and it won’t even be for AI to know when to use its skill set in one place and when to apply it elsewhere in a different context; so-called learning transfer, as Mike Loukides says. But rather, will AI reach the point where it cares?

  • Will AI ever care what it should know and where to stop or when it knows enough on any given subject?
  • How will it know or care if what it learns is true?
  • If in the best interests of advancing technology or through inaction  we do not limit its boundaries, what oversight is there of its implications?

Online limits will limit what we can reach in Thinking and Learning

If you look carefully at how humans learn online, I think rather than seeing  surprisingly little unsupervised learning, you see a lot of unsupervised questioning. It is often in the questioning that is done in private we discover, and through discovery we learn. Often valuable discoveries are made; whether in science, in maths, or important truths are found where there is a need to challenge the status quo. Imagine if Galileo had given up.

The freedom to think freely and to challenge authority, is vital to protect, and one reason why I and others are concerned about the compulsory web monitoring starting on September 5th in all schools in England, and its potential chilling effect. Some are concerned who  might have access to these monitoring results today or in future, if stored could they be opened to employers or academic institutions?

If you tell children do not use these search terms and do not be curious about *this* subject without repercussions, it is censorship. I find the idea bad enough for children, but for us as adults its scary.

As Frankie Boyle wrote last November, we need to consider what our internet history is:

“The legislation seems to view it as a list of actions, but it’s not. It’s a document that shows what we’re thinking about.”

Children think and act in ways that they may not as an adult. People also think and act differently in private and in public. It’s concerning that our private online activity will become visible to the State in the IP Bill — whether photographs that captured momentary actions in social media platforms without the possibility to erase them, or trails of transitive thinking via our web history — and third-parties may make covert judgements and conclusions about us, correctly or not, behind the scenes without transparency, oversight or recourse.

Children worry about lack of recourse and repercussions. So do I. Things done in passing, can take on a permanence they never had before and were never intended. If expert providers of the tech world such as Apple Inc, Facebook Inc, Google Inc, Microsoft Corp, Twitter Inc and Yahoo Inc are calling for change, why is the government not listening? This is more than very concerning, it will have disastrous implications for trust in the State, data use by others, self-censorship, and fear that it will lead to outright censorship of adults online too.

By narrowing our parameters what will we not discover? Not debate?  Or not invent? Happy are the clockmakers, and kids who create. Any restriction on freedom to access information, to challenge and question will restrict children’s learning or even their wanting to.  It will limit how we can improve our shared knowledge and improve our society as a result. The same is true of adults.

So in teaching AI how to learn, I wonder how the limitations that humans put on its scope — otherwise how would it learn what the developers want — combined with showing it ‘our thinking’ through search terms,  and how limitations on that if users self-censor due to surveillance, will shape what AI will help us with in future and will it be the things that could help the most people, the poorest people, or will it be people like those who programme the AI and use search terms and languages it already understands?

Who is accountable for the scope of what we allow AI to do or not? Who is accountable for what AI learns about us, from our behaviour data if it is used without our knowledge?

How far does AI have to go?

The leap for AI will be if and when AI can determine what it doesn’t know, and it sees a need to fill that gap. To do that, AI will need to discover a purpose for its own learning, indeed for its own being, and be able to do so without limitation from the that humans shaped its framework for doing so. How will AI know what it needs to know and why? How will it know, what it knows is right and sources to trust? Against what boundaries will AI decide what it should engage with in its learning, who from and why? Will it care? Why will it care? Will it find meaning in its reason for being? Why am I here?

We assume AI will know better. We need to care, if AI is going to.

How far are we away from a machine that is capable of recursive self-improvement, asks John Naughton in yesterday’s Guardian, referencing work by Yuval Harari suggesting artificial intelligence and genetic enhancements will usher in a world of inequality and powerful elites. As I was finishing this, I read his article, and found myself nodding, as I read the implications of new technology focus too much on technology and too little on society’s role in shaping it.

AI at the moment has a very broad meaning to the general public. Is it living with life-supporting humanoids?  Do we consider assistive search tools as AI? There is a fairly general understanding of “What is A.I., really?” Some wonder if we are “probably one of the last generations of Homo sapiens,” as we know it.

If the purpose of AI is to improve human lives, who defines improvement and who will that improvement serve? Is there a consensus on the direction AI should and should not take, and how far it should go? What will the global language be to speak AI?

As AI learning progresses, every time AI turns to ask its creators, “Are we there yet?”,  how will we know what to say?

image: Stephen Barling flickr.com/photos/cripsyduck (CC BY-NC 2.0)

#caredata, care.data, change, choice, confidentiality, datasharing, empowerment, ethics, HSCIC, National Pupil Database, NHS England, trust

Datasharing, lawmaking and ethics: power, practice and public policy

“Lawmaking is the Wire, not Schoolhouse Rock. It’s about blood and war and power, not evidence and argument and policy.”

"We can't trust the regulators," they say. "We need to be able to investigate the data for ourselves." Technology seems to provide the perfect solution. Just put it all online - people can go through the data while trusting no one.  There's just one problem. If you can't trust the regulators, what makes you think you can trust the data?"

Extracts from The Boy Who Could Change the World: The Writings of Aaron Swartz. Chapter: ‘When is Technology Useful? ‘ June 2009.

The question keeps getting asked, is the concept of ethics obsolete in Big Data?

I’ve come to some conclusions why ‘Big Data’ use keeps pushing the boundaries of what many people find acceptable, and yet the people doing the research, the regulators and lawmakers often express surprise at negative reactions. Some even express disdain for public opinion, dismissing it as ignorant, not ‘understanding the benefits’, yet to be convinced. I’ve decided why I think what is considered ‘ethical’ in data science does not meet public expectation.

It’s not about people.

Researchers using large datasets, often have a foundation in data science, applied computing, maths, and don’t see data as people. It’s only data. Creating patterns, correlations, and analysis of individual level data are not seen as research involving human subjects.

This is embodied in the nth number of research ethics reviews I have read in the last year in which the question is asked, does the research involve people? The answer given is invariably ‘no’.

And these data analysts using, let’s say health data, are not working in a subject that is founded on any ethical principle, contrasting with the medical world the data come from.

The public feels differently about the information that is about them, and may be known, only to them or select professionals. The values that we as the public attach to our data  and expectations of its handling may reflect the expectation we have of handling of us as people who are connected to it. We see our data as all about us.

The values that are therefore put on data, and on how it can and should be used, can be at odds with one another, the public perception is not reciprocated by the researchers. This may be especially true if researchers are using data which has been de-identified, although it may not be anonymous.

New legislation on the horizon, the Better Use of Data in Government,  intends to fill the [loop]hole between what was legal to share in the past and what some want to exploit today, and emphasises a gap in the uses of data by public interest, academic researchers, and uses by government actors. The first incorporate by-and-large privacy and anonymisation techniques by design, versus the second designed for applied use of identifiable data.

Government departments and public bodies want to identify and track people who are somehow misaligned with the values of the system; either through fraud, debt, Troubled Families, or owing Student Loans. All highly sensitive subjects. But their ethical data science framework will not treat them as individuals, but only as data subjects. Or as groups who share certain characteristics.

The system again intrinsically fails to see these uses of data as being about individuals, but sees them as categories of people – “fraud” “debt” “Troubled families.” It is designed to profile people.

Services that weren’t built for people, but for government processes, result in datasets used in research, that aren’t well designed for research. So we now see attempts to shoehorn historical practices into data use  by modern data science practitioners, with policy that is shortsighted.

We can’t afford for these things to be so off axis, if civil service thinking is exploring “potential game-changers such as virtual reality for citizens in the autism spectrum, biometrics to reduce fraud, and data science and machine-learning to automate decisions.”

In an organisation such as DWP this must be really well designed since “the scale at which we operate is unprecedented: with 800 locations and 85,000  colleagues, we’re larger than most retail operations.”

The power to affect individual lives through poor technology is vast and some impacts seem to be being badly ignored. The ‘‘real time earnings’ database improved accuracy of benefit payments was widely agreed to have been harmful to some individuals through the Universal Credit scheme, with delayed payments meaning families at foodbanks, and contributing to worse.

“We believe execution is the major job of every business leader,” perhaps not the best wording in on DWP data uses.

What accountability will be built-by design?

I’ve been thinking recently about drawing a social ecological model of personal data empowerment or control. Thinking about visualisation of wants, gaps and consent models, to show rather than tell policy makers where these gaps exist in public perception and expectations, policy and practice. If anyone knows of one on data, please shout. I think it might be helpful.

But the data *is* all about people

Regardless whether they are in front of you or numbers on a screen, big or small datasets using data about real lives are data about people. And that triggers a need to treat the data with an ethical approach as you would people involved face-to-face.

Researchers need to stop treating data about people as meaningless data because that’s not how people think about their own data being used. Not only that, but if the whole point of your big data research is to have impact, your data outcomes, will change lives.

Tosh, I know some say. But, I have argued, the reason being is that the applications of the data science/ research/ policy findings / impact of immigration in education review / [insert purposes of the data user’s choosing] are designed to have impact on people. Often the people about whom the research is done without their knowledge or consent. And while most people say that is OK, where it’s public interest research, the possibilities are outstripping what the public has expressed as acceptable, and few seem to care.

Evidence from public engagement and ethics all say, hidden pigeon-holing, profiling, is unacceptable. Data Protection law has special requirements for it, on autonomous decisions. ‘Profiling’ is now clearly defined under article 4 of the GDPR as ” any form of automated processing of personal data consisting of using those 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.”

Using big datasets for research that ‘isn’t interested in individuals’ may still intend to create results profiling groups for applied policing, or discriminate, to make knowledge available by location. The data may have been deidentified, but in application becomes no longer anonymous.

Big Data research that results in profiling groups with the intent for applied health policy impacts for good, may by the very point of research, with the intent of improving a particular ethnic minority access to services, for example.

Then look at the voting process changes in North Carolina and see how that same data, the same research knowledge might be applied to exclude, to restrict rights, and to disempower.

Is it possible to have ethical oversight that can protect good data use and protect people’s rights if they conflict with the policy purposes?

The “clear legal basis”is not enough for public trust

Data use can be legal and can still be unethical, harmful and shortsighted in many ways, for both the impacts on research – in terms of withholding data and falsifying data and avoiding the system to avoid giving in data – and the lives it will touch.

What education has to learn from health is whether it will permit the uses by ‘others’ outside education to jeopardise the collection of school data intended in the best interests of children, not the system. In England it must start to analyse what is needed vs wanted. What is necessary and proportionate and justifies maintaining named data indefinitely, exposed to changing scope.

In health, the most recent Caldicott review suggests scope change by design – that is a red line for many: “For that reason the Review recommends that, in due course, the opt-out should not apply to all flows of information into the HSCIC. This requires careful consideration with the primary care community.”

The community spoke out already, and strongly in Spring and Summer 2014 that there must be an absolute right to confidentiality to protect patients’ trust in the system. Scope that ‘sounds’ like it might sneakily change in future, will be a death knell to public interest research, because repeated trust erosion will be fatal.

Laws change to allow scope change without informing people whose data are being used for different purposes

Regulators must be seen to be trusted, if the data they regulate is to be trustworthy. Laws and regulators that plan scope for the future watering down of public protection, water down public trust from today. Unethical policy and practice, will not be saved by pseudo-data-science ethics.

Will those decisions in private political rooms be worth the public cost to research, to policy, and to the lives it will ultimately affect?

What happens when the ethical black holes in policy, lawmaking and practice collide?

At the last UK HealthCamp towards the end of the day, when we discussed the hard things, the topic inevitably moved swiftly to consent, to building big databases, public perception, and why anyone would think there is potential for abuse, when clearly the intended use is good.

The answer came back from one of the participants, “OK now it’s the time to say. Because, Nazis.” Meaning, let’s learn from history.

Given the state of UK politics, Go Home van policies, restaurant raids, the possibility of Trump getting access to UK sensitive data of all sorts from across the Atlantic, given recent policy effects on the rights of the disabled and others, I wonder if we would hear the gentle laughter in the room in answer to the same question today.

With what is reported as Whitehall’s digital leadership sharp change today, the future of digital in government services and policy and lawmaking does indeed seem to be more “about blood and war and power,” than “evidence and argument and policy“.

The concept of ethics in datasharing using public data in the UK is far from becoming obsolete. It has yet to begin.

We have ethical black holes in big data research, in big data policy, and big data practices in England. The conflicts between public interest research and government uses of population wide datasets, how the public perceive the use of our data and how they are used, gaps and tensions in policy and practice are there.

We are simply waiting for the Big Bang. Whether it will be creative, or destructive we are yet to feel.

*****

image credit: LIGO – graphical visualisation of black holes on the discovery of gravitational waves

References:

Report: Caldicott review – National Data Guardian for Health and Care Review of Data Security, Consent and Opt-Outs 2016

Report: The One–Way Mirror: Public attitudes to commercial access to health data

Royal Statistical Society Survey carried out by Ipsos MORI: The Data Trust Deficit

comm consent, communications, confidentiality, consent, datasharing, engagement, ethics, politics, privacy, research, surveillance, Tech, transparency, trust

The illusion that might cheat us: ethical data science vision and practice

This blog post is also available as an audio file on soundcloud.

Anais Nin, wrote in her 1946 diary of the dangers she saw in the growth of technology to expand our potential for connectivity through machines, but diminish our genuine connectedness as people. She could hardly have been more contemporary for today:

“This is the illusion that might cheat us of being in touch deeply with the one breathing next to us. The dangerous time when mechanical voices, radios, telephone, take the place of human intimacies, and the concept of being in touch with millions brings a greater and greater poverty in intimacy and human vision.” [Extract from volume IV 1944-1947]

Echoes from over 70 years ago, can be heard in the more recent comments of entrepreneur Elon Musk. Both are concerned with simulation, a lack of connection between the perceived, and reality, and the jeopardy this presents for humanity. But both also have a dream. A dream based on the positive potential society has.

How will we use our potential?

Data is the connection we all have between us as humans and what machines and their masters know about us. The values that masters underpin their machine design with, will determine the effect the machines and knowledge they deliver, have on society.

In seeking ever greater personalisation, a wider dragnet of data is putting together ever more detailed pieces of information about an individual person. At the same time data science is becoming ever more impersonal in how we treat people as individuals. We risk losing sight of how we respect and treat the very people whom the work should benefit.

Nin grasped the risk that a wider reach, can mean more superficial depth. Facebook might be a model today for the large circle of friends you might gather, but how few you trust with confidences, with personal knowledge about your own personal life, and the privilege it is when someone chooses to entrust that knowledge to you. Machine data mining increasingly tries to get an understanding of depth, and may also add new layers of meaning through profiling, comparing our characteristics with others in risk stratification.
Data science, research using data, is often talked about as if it is something separate from using information from individual people. Yet it is all about exploiting those confidences.

Today as the reach has grown in what is possible for a few people in institutions to gather about most people in the public, whether in scientific research, or in surveillance of different kinds, we hear experts repeatedly talk of the risk of losing the valuable part, the knowledge, the insights that benefit us as society if we can act upon them.

We might know more, but do we know any better? To use a well known quote from her contemporary, T S Eliot, ‘Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?’

What can humans achieve? We don’t yet know our own limits. What don’t we yet know?  We have future priorities we aren’t yet aware of.

To be able to explore the best of what Nin saw as ‘human vision’ and Musk sees in technology, the benefits we have from our connectivity; our collaboration, shared learning; need to be driven with an element of humility, accepting values that shape  boundaries of what we should do, while constantly evolving with what we could do.

The essence of this applied risk is that technology could harm you, more than it helps you. How do we avoid this and develop instead the best of what human vision makes possible? Can we also exceed our own expectations of today, to advance in moral progress?

Continue reading The illusion that might cheat us: ethical data science vision and practice →

care.data, change management, communications, engagement, ethics, genomics, health, human rights, politics, privacy, research, transparency, trust

OkCupid and Google DeepMind: Happily ever after? Purposes and ethics in datasharing

This blog post is also available as an audio file on soundcloud.

What constitutes the public interest must be set in a universally fair and transparent ethics framework if the benefits of research are to be realised – whether in social science, health, education and more – that framework will provide a strategy to getting the pre-requisite success factors right, ensuring research in the public interest is not only fit for the future, but thrives. There has been a climate change in consent. We need to stop talking about barriers that prevent datasharing  and start talking about the boundaries within which we can.

What is the purpose for which I provide my personal data?

‘We use math to get you dates’, says OkCupid’s tagline.

That’s the purpose of the site. It’s the reason people log in and create a profile, enter their personal data and post it online for others who are looking for dates to see. The purpose, is to get a date.

When over 68K OkCupid users registered for the site to find dates, they didn’t sign up to have their identifiable data used and published in ‘a very large dataset’ and onwardly re-used by anyone with unregistered access. The users data were extracted “without the express prior consent of the user […].”

Are the registration consent purposes compatible with the purposes to which the researcher put the data should be a simple enough question.  Are the research purposes what the person signed up to, or would they be surprised to find out their data were used like this?

Questions the “OkCupid data snatcher”, now self-confessed ‘non-academic’ researcher, thought unimportant to consider.

But it appears in the last month, he has been in good company.

Google DeepMind, and the Royal Free, big players who do know how to handle data and consent well, paid too little attention to the very same question of purposes.

The boundaries of how the users of OkCupid had chosen to reveal information and to whom, have not been respected in this project.

Nor were these boundaries respected by the Royal Free London trust that gave out patient data for use by Google DeepMind with changing explanations, without clear purposes or permission.

The legal boundaries in these recent stories appear unclear or to have been ignored. The privacy boundaries deemed irrelevant. Regulatory oversight lacking.

The respectful ethical boundaries of consent to purposes, disregarding autonomy, have indisputably broken down, whether by commercial org, public body, or lone ‘researcher’.

Research purposes

The crux of data access decisions is purposes. What question is the research to address – what is the purpose for which the data will be used? The intent by Kirkegaard was to test:

“the relationship of cognitive ability to religious beliefs and political interest/participation…”

In this case the question appears intended rather a test of the data, not the data opened up to answer the test. While methodological studies matter, given the care and attention [or self-stated lack thereof] given to its extraction and any attempt to be representative and fair, it would appear this is not the point of this study either.

The data doesn’t include profiles identified as heterosexual male, because ‘the scraper was’. It is also unknown how many users hide their profiles, “so the 99.7% figure [identifying as binary male or female] should be cautiously interpreted.”

“Furthermore, due to the way we sampled the data from the site, it is not even representative of the users on the site, because users who answered more questions are overrepresented.” [sic]

The paper goes on to say photos were not gathered because they would have taken up a lot of storage space and could be done in a future scraping, and

“other data were not collected because we forgot to include them in the scraper.”

The data are knowingly of poor quality, inaccurate and incomplete. The project cannot be repeated as ‘the scraping tool no longer works’. There is an unclear ethical or peer review process, and the research purpose is at best unclear. We can certainly give someone the benefit of the doubt and say intent appears to have been entirely benevolent. It’s not clear what the intent was. I think it is clearly misplaced and foolish, but not malevolent.

The trouble is, it’s not enough to say, “don’t be evil.” These actions have consequences.

When the researcher asserts in his paper that, “the lack of data sharing probably slows down the progress of science immensely because other researchers would use the data if they could,”  in part he is right.

Google and the Royal Free have tried more eloquently to say the same thing. It’s not research, it’s direct care, in effect, ignore that people are no longer our patients and we’re using historical data without re-consent. We know what we’re doing, we’re the good guys.

However the principles are the same, whether it’s a lone project or global giant. And they’re both wildly wrong as well. More people must take this on board. It’s the reason the public interest needs the Dame Fiona Caldicott review published sooner rather than later.

Just because there is a boundary to data sharing in place, does not mean it is a barrier to be ignored or overcome. Like the registration step to the OkCupid site, consent and the right to opt out of medical research in England and Wales is there for a reason.

We’re desperate to build public trust in UK research right now. So to assert that the lack of data sharing probably slows down the progress of science is misplaced, when it is getting ‘sharing’ wrong, that caused the lack of trust in the first place and harms research.

A climate change in consent

There has been a climate change in public attitude to consent since care.data, clouded by the smoke and mirrors of state surveillance. It cannot be ignored.  The EUGDPR supports it. Researchers may not like change, but there needs to be an according adjustment in expectations and practice.

Without change, there will be no change. Public trust is low. As technology advances and if we continue to see commercial companies get this wrong, we will continue to see public trust falter unless broken things get fixed. Change is possible for the better. But it has to come from companies, institutions, and people within them.

Like climate change, you may deny it if you choose to. But some things are inevitable and unavoidably true.

There is strong support for public interest research but that is not to be taken for granted. Public bodies should defend research from being sunk by commercial misappropriation if they want to future-proof public interest research.

The purpose for which the people gave consent are the boundaries within which you have permission to use data, that gives you freedom within its limits, to use the data.  Purposes and consent are not barriers to be overcome.

If research is to win back public trust developing a future proofed, robust ethical framework for data science must be a priority today.

Commercial companies must overcome the low levels of public trust they have generated in the public to date if they ask ‘trust us because we’re not evil‘. If you can’t rule out the use of data for other purposes, it’s not helping. If you delay independent oversight it’s not helping.

This case study and indeed the Google DeepMind recent episode by contrast demonstrate the urgency with which working out what common expectations and oversight of applied ethics in research, who gets to decide what is ‘in the public interest’ and data science public engagement must be made a priority, in the UK and beyond.

Boundaries in the best interest of the subject and the user

Society needs research in the public interest. We need good decisions made on what will be funded and what will not be. What will influence public policy and where needs attention for change.

To do this ethically, we all need to agree what is fair use of personal data, when is it closed and when is it open, what is direct and what are secondary uses, and how advances in technology are used when they present both opportunities for benefit or risks to harm to individuals, to society and to research as a whole.

The potential benefits of research are potentially being compromised for the sake of arrogance, greed, or misjudgement, no matter intent. Those benefits cannot come at any cost, or disregard public concern, or the price will be trust in all research itself.

In discussing this with social science and medical researchers, I realise not everyone agrees. For some, using deidentified data in trusted third party settings poses such a low privacy risk, that they feel the public should have no say in whether their data are used in research as long it’s ‘in the public interest’.

For the DeepMind researchers and Royal Free, they were confident even using identifiable data, this is the “right” thing to do, without consent.

For the Cabinet Office datasharing consultation, the parts that will open up national registries, share identifiable data more widely and with commercial companies, they are convinced it is all the “right” thing to do, without consent.

How can researchers, society and government understand what is good ethics of data science, as technology permits ever more invasive or covert data mining and the current approach is desperately outdated?

Who decides where those boundaries lie?

“It’s research Jim, but not as we know it.” This is one aspect of data use that ethical reviewers will need to deal with, as we advance the debate on data science in the UK. Whether independents or commercial organisations. Google said their work was not research. Is‘OkCupid’ research?

If this research and data publication proves anything at all, and can offer lessons to learn from, it is perhaps these three things:

Who is accredited as a researcher or ‘prescribed person’ matters. If we are considering new datasharing legislation, and for example, who the UK government is granting access to millions of children’s personal data today. Your idea of a ‘prescribed person’ may not be the same as the rest of the public’s.

Researchers and ethics committees need to adjust to the climate change of public consent. Purposes must be respected in research particularly when sharing sensitive, identifiable data, and there should be no assumptions made that differ from the original purposes when users give consent.

Data ethics and laws are desperately behind data science technology. Governments, institutions, civil, and all society needs to reach a common vision and leadership how to manage these challenges. Who defines these boundaries that matter?

How do we move forward towards better use of data?

Our data and technology are taking on a life of their own, in space which is another frontier, and in time, as data gathered in the past might be used for quite different purposes today.

The public are being left behind in the game-changing decisions made by those who deem they know best about the world we want to live in. We need a say in what shape society wants that to take, particularly for our children as it is their future we are deciding now.

How about an ethical framework for datasharing that supports a transparent public interest, which tries to build a little kinder, less discriminating, more just world, where hope is stronger than fear?

Working with people, with consent, with public support and transparent oversight shouldn’t be too much to ask. Perhaps it is naive, but I believe that with an independent ethical driver behind good decision-making, we could get closer to datasharing like that.

That would bring Better use of data in government.

Purposes and consent are not barriers to be overcome. Within these, shaped by a strong ethical framework, good data sharing practices can tackle some of the real challenges that hinder ‘good use of data’: training, understanding data protection law, communications, accountability and intra-organisational trust. More data sharing alone won’t fix these structural weaknesses in current UK datasharing which are our really tough barriers to good practice.

How our public data will be used in the public interest will not be a destination or have a well defined happy ending, but it is a long term  process which needs to be consensual and there needs to be a clear path to setting out together and achieving collaborative solutions.

While we are all different, I believe that society shares for the most part, commonalities in what we accept as good, and fair, and what we believe is important. The family sitting next to me have just counted out their money and bought an ice cream to share, and the staff gave them two. The little girl is beaming. It seems that even when things are difficult, there is always hope things can be better. And there is always love.

Even if some might give it a bad name.

********

img credit: flickr/sofi01/ Beauty and The Beast  under creative commons