miércoles, 21 de diciembre de 2016

Free Flow of Data in Europe: The story of the cake

The European Commission has delivered almost on time the majority the legislative and non-legislative proposals contained in the Digital Single Market Strategy. Some of them even are expected to be approved by the Council and Parliament by the first half of 2017. However, there is one of them that looks trapped in a blind alley: the free flow data initiative (FFDI).

The FFDI was initially expected to be presented in september 2016, later was delayed to november 2016 after the High Level Conference on European Data Economy, and now it is not expected until the end of 2017 Q1. In spite of the pressure of part of the industry and the sympathies of some governments, there is growing doubt about its future existence.

There is war on figures about the economic impact of the liberalisation of the free flow of data, particularly the establishment of a ban on data localisation restrictions. In a manifesto following the last Telecomunnications Council, the industry estimated the potential of data in Europe in 566 EUR billion by 2020. A more realistic estimation of ECIPE put on the table the figure of 8 EUR billion per year (which means much less than 1% of EU GDP).

But free flow of data is beyond economics, it is about building Europe. I'm greatly agree with the need to promote free flow of data in Europe and the need of regulations for its implementation. After the achievement of the freedom of movement of people, products, services and capital, the free flow of data is the 5th freedom that should define the European Unión in the digital era. However, free flow of data should not be restricted to build a regulation for deregulated massively data location requirements. The regulation of  data ownership, access and liability it is also needed to ensure legal certainty in the flow of data. 

You can't have your cake and eat it. The debate around the ownership and access to data is central for the future of our society. As Morozov remembered us recently, five firms alone digest the majority of the world data. Imagine a present where five companies own the property of land in the world to foreseen the consequences for the future if we don´t regulate the access and ownership to data right now.

Free flow of data is needed, but goes beyond the elimination of data requirements. You can´t have a cake and eat it at the same time.


miércoles, 14 de diciembre de 2016

The disturbing investigation of @propublica on algorithms

One of the more interesting web sites concerning the algorithmic society that I have discover along the last months is Propublica. It is not a website specialised on the digital economy but in investigative journalism in the public interest. In the last months they have conducted an in-depth investigation on decisions taken by algorithms and the consequences resulted from them. What follows it is just some of the cases I have found there, enough for strengthening my conviction that regulators should do something more on the issue than be just reactive.

The price of items and services sold in websites is the focus of some of the cases. Due to its dominance on e-commerce, you may be interested in reading the article about the algorithm used by Amazon to present the answer to your requests about any object that you desire to buy. Propublica has discovered how the presentation of the results ordered by price has some tricks that favoured the items sold directly by Amazon. Other article shows Propublica´s discoveries around the different pricing based on geographical area that the Pricenton Review charges for their online courses. Do Consumer Authorities have the capabilities to make such discoveries?

Let´s go to principles. I assume the majority of the people who read this blog are strongly against any kind of racial discrimination. Probably, if you read the articles regarding algorithmic racial discrimination in Propublica you would feel sick. On one hand, there are evidences that algorithms used by public authorities to predict crime in USA are heavily biased against afroamericans. On the other hand, they show how facebook use the racial classification for sorting advertisements. Any of us have any cause to be discriminated, and data hoarders are able to discriminate you. Propublica also discovered that Facebook offered to companies more than 1300 ad categories.

The most disturbing feeling you have after reading the results of Propublica research is that you would feel that you are facing just to the tip of the iceberg. Someone has dedicated time and efforts and have made some discoveries. But what have yet to be known about how the companies handle our digital profiles?


miércoles, 7 de diciembre de 2016

Regulating the objects: Stopping Orwell

The Internet of the Things is not anymore a distant concept but a daily reality. Undoubtedly, the prediction of 20 billion of objects connected by 2020 is impressive, but equally impressive  it is the estimation of nearly 5 billion connected right now. And the main driver of this rise of connected object will not be production environments as the Smart Industry, but consumers. Both in the short and long term more than half of the connected objects will be plugged by consumers.

Almost everyday we can read pieces of news regarding new kind of smart objects. For instance, in less than a week The Guardian published an article regarding the impressive growth of the smart toys market and a piece of new about digitalised sex toys. It doesn´t matter if we feel a little bit akward reading this kind texts, we will not be able to stop the introduction of chips and conectivity  capabilities in all the object surrounding us. Or were we capable in the past of stopping the disapearance of the TV with cathode ray tube?

So instead of refraining of the regulation intentions for the sake of innovation and ask for a light touch approach, we should encourage governments to explore and develop all the needed regulatory options. As any kind of objects are different, we would certainly need specific regulations  for classes of objects instead of general ones. However, we should not overlook the need for transversal rules defining common principles.  An example of the former is the US checklist for self-driving cars.  Among the latter we can include the  "Guide to the ethical design and application of robots and robotic systems" developed by the British Institute for Standards.

Specially important is the regulation of the issues regarding the handling of data privacy and data sharing. Not in vain, both of them were at the top of the US checklist for self-driving cars.  The "purpose limitation principle" and the "data minimisation" principles should be at the centre of data regulation for Internet of the Things.

1984 has arrived to our lives more than 30 years later. Not only the TV has surveillance capabilities, but all the objects around. Stopping the Orwell´s nightmare before it takes form depends on the rules we establish today.
palyginti kainas