The subversion of democracy by big data

The subversion of democracy by big data

The fabulous Carol Cadwalladyr brings us the next instalment of undoing the surveillance states control over our democracies.

In an article “The Great British Brexit Robbery”, she and the Guardian showed how the Tories and the Brexit Leave Campaigns had used US Data Aanlytics companies to influence the Brexit referendum. It is alleged that the personal data was obtained illegally, its processing was illegal and that it was an undeclared election/referendum expense. The evidence was sufficient for the Information Commissioner’s Office and the Electoral Commission to launch investigations.

Over the last two days, Facebook have suspended Cambridge Analytica & one other company and the latter’s Principal for breaking their terms and conditions and in one case a breach of contract not to pass data on. The story is reported in the Guardian in a story called, “‘I made Steve Bannon’s psychological warfare tool’: meet the data war whistleblower” , which documents the contractual paper trial. This happened two years ago and it is alleged that Facebook knew of it then. It is a crime in many jurisdictions, including California to not notify either the regulators or the data subjects of a breach/leak of personal data.

Sadly 🤔 they have been accused of misleading the House of Commons, select committee inquiry into Fake News. It has been denied that Cambridge Analytica had Facebook data in a verbal submission. Its Chair, Damian Collins, is quite forthright, accusing Facebook of sending under informed representatives to answer the committee’s questions. The word wilful ignorance comes to mind.

As Brits, we need to see if crimes were committed during the 2015 & 2017 General Elections and/or the Brexit Refrendum but this can’t be good for Facebook’s reputation.


I wish we still had Storify, this is one for them.

The image is from the Guardian on the story on Parliament’s reaction. …


Hannes Grassegger and Mikael Krogerus explore the latest in Election winning software as sold by Cambridge Analytica and used by the Republican Party in the US General Election. It was also seemingly used by the Brexit campaigners in the UK. Their article explores the evolution of the pyschograph model and data collection techniques developed by academics and then looks at its more sinister uses via Facebook profiling data and big data analytics.

There are two lessons from this. The first, easier and personal one is to seriously consider how one uses Facebook, if at all. I feel it harder to give up now that messenger has E2E encryption but I shan’t be doing any more quizzes, not even what D&D class would you be;  I suppose especially not what D&D class are you, (Wizard by the way). Even the fact that I did an ABBA quiz yesterday and not Lady Gaga seems to give these nosy bastards data that they use.

The second lesson is for the Labour Party and relates to the fact that Trump’s campaign (and to some extent the Brexit and UK GE 2015) have updated the IT playbook. Up until then Obama’s 2012 campaign was the model to match and beat. To compare with this, in 2015, Labour spent more on the “Ed Stone” detailing the infamous five pledges than on Facebook and while some message prototyping, where messages are tested for effectiveness and then targeted and a tiny amount of crowd sourced message design occurred, we have to do better.

In the run up to the election, both Labour and the Tories brought in consultants from the Obama campaign, Labour brought in David Axelrod, and the Tories brought in Jim Messina. Messina would seem to have been the big data champion in Obama’s campaign and with the help of Cambridge Analytica they rewrote the political IT playbook.

The Electoral Commission are now looking into the legality of the Tories expenditure; micro-targeting and message customisation has been shown to cause campaign accounting problems since local and national expenses need to be accounted for separately and have separate limits.

And this is all before the authorities consider the data protection and privacy implications.

There are dangers from micro-targeting as promises made to some groups can be used against the authors. The Motherboard article highlights Clinton’s problem over relief in Haiti and in the UK, Ed Miliband’s mansion tax did not play well in London and the way the British deal with racism in politics has always given hostages to fortune, as we can see on some of the attacks on Ken Livingstone and aspects of the 2016 London Mayoral campaign.

The use of pyschographs and the messages it creates is a reinforcement of the the move from a transactional politics to perception based campaign; it’s one of the reasons that Ed Miliband failed and Labour’s new left, but aged leadership can’t break through to the electorate and one of the reasons why May can adopt elements of Miliband’s programme on energy and the minimum wage. Programmes and promises don’t count as much as they used to.

How to change perception is hard. I have been struggling with “Perception is reality” for many years because it isn’t. There are facts beyond perception and ignorance is ignorance.

Fixing Labour’s IT is easier, and the answer isn’t PHP and an updated voter ID system, it needs a modern applications architecture, a big data platform and it would seem they need to resurrect the once famed rapid rebuttal system. The excuse we can’t afford it no longer holds true. Did you know Labour is debt free?


This is a longer comment based on what I said facebook yesterday, I have made a story @ storify which has some further sources. This article was published on, it has tool tips and should probably have been posted on the blog. …

Privacy and Big Data

I read Privacy and Big Data by Craig and Ludloff towards the end of 2013. The first chapter is called “The Perfect Storm”. The book lists a number of consumer and corporate computing trends, from Google’s search solution and their clustered file systems, the consumer adoption of cloud storage and the realisation of parallel computing models. There is no question that data is growing at an explosive rate and that new computational models are being developed to use these new volumes of data in timescales appropriate to the human. These new models are of interest to both the new internet companies and to Governments yet because of both social media and the distributed nature of modern computing raise questions of privacy. …

New problems, new architectures

I originally wrote a piece on my sun/oracle blog called “Are “Quants” suitable for grid infrastructure?” I used a finance dictionary to create a dichotomy between Quants & Chartists and suggested that new IT architectures were needed to deal with trends and forecasting, especially if applied to portfolios and baskets. I reproduce the words here, as I was clearly struggling towards a forecast about big data and new application’s and implementation architectures. The original is not tagged “big data” as it wasn’t a current term at that time. …