miércoles, 13 de marzo de 2019

Huawei War in movie mode

I am totally sure that in a few years time the controversy around 5G security will be remember among the first technology battles between USA and China, but surely not the last one. The time will say if we will remember the controversy as an equivalent to the Cuban´s Missile Crisis  or to the Weapons of Mass Destruction theory that paved the way to the 2nd Irak War. Or what it is the same,  as an episode where the US Administration provided evidences of a great conspirancy against world´s peace of a rivalry country or as the day that a  forgettable US  President tried to drive the whole world in a war without any real evidence.

We can compare this war with past wars and tactics, which give us a better view of the main actors. As i have already mentioned, it is difficult to say if US or Huawei are on the right side of the war, but you can find reasons to suspect from both.

To begin with US, we can start with the incident that start openly this war. In the manner of the I World War, a Sarajevo alike incident was suffered by an ally (the arrest of Huawei´s founder by Canada) and after that US started to scale-up the war. In this war, as what happened in the 2nd Irak War, US is pressing their European allies looking for a domino effect. So far without success, but another coincidence with the 2nd Irak War is that the sole declaration showing concerns about Huawei has came from Commissioner Ansip (an Estonian), part of the so-called new Europe.

Huawei for its part has not been waiting stoically the blows. Clearly, its tactics has been oriented to avoid the domino effect in Europe. And the main focus of its actions has been UK. Maybe, watching the interview with Huawei´s owner and the reference to their investments in UK (which are really huge), someone has remembered the scenes of the Godfather where Brando remember to someone all the things he has done for him. Probably, it is coincidence that the cautious position of UK on banning Huawei has been heavily disseminated around the same days.

The war is still going on. We will have yet to see some more battles of it. It is difficult to say if we will know who is right or wrong anytime soon. However, we what is sure is that we will still have to watch interesting episodes, it is advisable to have at hand some popcorns and your favourite soft drink.

miércoles, 6 de marzo de 2019

Algorithm transparency regulation

Europe has become the greatest Silicon´s Valley regulator. The Digital Single Market Strategy which started its implementation in 2015 has contributed to slash the fragmentation of the digital rules in Europe, and this achievement has been reached mainly through the definition of common rules for online businesses. As the main digital businesses in the world are located in Silicon´s Valley, the main outcome of the Digital Single Market is thirty new regulations which will have impact in the GAFA´s business models. 

And the impact of the new EU regulations will be global, not reduced to Europe. Now, the GAFA´s could threat with the abandonment of EU markets and how the rules stifle innovation, but its the result of their success in their lobby activities in Washington. As Washington has renounced to the regulation of the GAFAs, the EU has filled the void in a manner less beneficious to their interests. This is a lesson the have learned and since one year ago we have seen a less aggressive approach to regulation form the GAFAs: You cannot avoid indefinitely regulation so it is better to lobby for a friendly one than for not having any one at all.

One of the main worries of firms, consumers and citizens is the transparency of the algorithms at the core of GAFAs and other online platforms. There's not a neo luddite movement strong enough which directly calls for not using platforms and the platform dividend is recognised by its users. However, there´s a strong desire for the tools that allow a bigger knowledge on their algorithms and a validation of its fair and ethical design.

The Digital Single Market has made some progress also in the area of regulated algorithm transparency. On one hand, the right to explanation included in the Article 22 of the GDPR obligues platform to have at hand an explanation on how personal data affects the conclusion achieved by an algorithm. On the other hand, the P2B relationships regulation introduces the obligation to disclose the main parameters the platforms use to rank goods and services on their site.

However, there are still many other kind of algorithms not cover by the obligations of transparency included in the above EU regulations. Some of them as important as the work allocation algorithms used by Uber or Mechanical Turk, for instance. Even if we manage to regulate a new handful of types there would be more kind of algorithms unregulated than regulated. Therefore, the solution should be a flexible and holistic regulation, but soft enough for not stifling innovation.

One possible solution is taken the approach follow for the regulation of the pharmaceutical industry, with the creation of protocols and guidelines for the development of algorithms and agencies that may analise their design and the fulfillment of them within the respect of the commercial secret. The debate may be if a unique agency would be enough for the regulation of such as transversal matter as algorithms or we would need specialised agencies for each economic sector.

But all of this maybe the work for the renewal of the renewal of the digital single market strtaegy.

miércoles, 20 de febrero de 2019

AI and SDG in least developed countries

In 2015, the United Nations published the 17 Sustainable Development Goals (SDGs) as a roadmap to guarantee a sustainable future by 2030. These goals range from ending hunger and poverty over realizing sustainable energy and gender equality to preserving our biodiversity. The tech community received the SDGs with a certain discomfort because, after an intense debate, there were not a specific goal for technological development.

In spite of the disappearance of technology from the front line of the SDGs, connectivity was immediately seen as a key driver for achieving the SDGs. Furthermore, increasing access to internet  and the proportion of global population covered by a mobile network were among the indicators and targets for the 9th SDG (9.c & 9.c.1). But until now, no other kind of technology different to connectivity were linked with achieving the SDGs. 

Everything has changed with the recent rise of interest for AI. Besides the debates on AI and ethics or AI and jobs which are happened in developed countries, ITU and UN organisations are pushing the debate on "AI for Good". The aim is to identify impactful AI solutions able to yield long-term benefits and help achieve the Sustainable Development Goals. And there is not a lack of them. AI applications focused in healthcare, energy, agriculture, education and environmental protection. are blooming.

Undoubtedly, the usage of AI technology is surrounded by fears. We demand explainability and robustness of AI, we demand and ethical approach to AI. But the approach is completely different when you are fighting malnutrition or pandemias. In a more pragmatic view, their users claim that "We don't really need to understand these systems fully so long as the outcomes are good." Perhaps, it ís the only possible approach for the usage of AI to the achievement of SDGs in least developed countries. However, in the production side of the AI systems in the developed countries we have the obligation to guarantee this goodness.  As citizens of developed countries, we should be vigilant that AI is not used by our government as a tool for neo-colonialism taking advantage of the needs of the Least Developed Countries.

miércoles, 13 de febrero de 2019

"Berta Isla" - Javier Marias

Berta IslaBerta Isla by Javier Marías
My rating: 4 of 5 stars

Javier Marias fiel a su estilo y a sus fantasmas. "Berta Isla" es una continuación de su escritura hipnótico y ultra descriptiva que nos lleva a explorar de su pluma todas las posibilidades de una escena, un retorno a los escenarios de fondo conectados con las temáticas mas oscuras del mundo británico y las citas literarias. Y, como siempre, en primer plano los matices de las relaciones de pareja, plasmando en situaciones extremas lo que son las situaciones cotidianas del mundo compartido entre dos.

El centro de "Berta Isla" es el conocimiento imposible de quién convive con nosotros. Vemos del otro lo que nos enseña, lo que nos cuenta, pero siempre detrás puede haber un universo que desconozcamos. Tanto Berta, como, en menor medida, su marido Tomás se desconocen el uno al otro. Y la paradoja es cómo ese desconocimiento va en aumento a partir de que su relación se consolida en un proyecto de futuro. Sin embargo, ese desconocimiento creciente no impide que el amor entre ellos crezca, incluso en la separación eterna y la espera de un reencuentro que es el segundo foco del libro.

"Berta Isla" es, también, una actualización del mito de Penélope y Ulises. En la novela, Javier Marias bebe los vientos homéricos y construye una historia de espera femenina, magníficamente plasmada en la foto de su portada. La resistencia a creer en la muerte del ser amado aunque todo señale lo contrario, pero al mismo tiempo, el deber de seguir viviendo, seguir progresando en una vida independiente del otro.

"Berta Isla" es una novela que podría ser trasladada a cualquier escenario de fondo. Pero ahí también Javier Marias realiza una construcción perfecta. La conexión con una de sus anteriores novelas, "Tu rostro mañana", dota a "Berta Isla" del trasfondo del mundo del espionaje británico, incluida su relación con el mundo académico de Oxford tan querido para el autor.

Como siempre que uno acaba de leer una novela de Javier Marias, varías preguntas quedan en el aire. La primera, porqué no leí esta novela inmediatamente cuando fue publicada. La segunda, cuánto tiempo tardará en publicar Javier Marias su siguiente obra maestra. La tercera, porqué hay quien piensa que Javier Marías no se merece el Premio Nobel.

View all my reviews

miércoles, 6 de febrero de 2019

The dangerous path to AI based robo politicians

From sex to the elderly care sector and tv news anchors, news of AI-based robots able of taking by assault the jobs of the more diverse sectors important for human relationships catch our attention and capture our imagination. This kind of AI-based robot applications transport us to a future where humans may spend a relevant part of their days being satisfied in their need of a companion by not human beings. One very particular kind of this robot applications are politicians.

As politicians take decisions and develop actions that have the biggest influence in our daily life, the idea of being on the hands  of an extremely efficient and capable "being " maybe highly attractive. The idea of 24 hours, 7 days a weeks politician, always available to explain the decisions he/she/it takes on behalf of us, decisions which are based on evidence without spurious influences, is capable to attract many people. Furthermore, the debate is open in prestigious media such as The Economist. What many years ago would have been taken as a freaky debate has now been introduced in the political agenda.

However, it is surprising that the issue even can be seriously considered. There are well known the cases of the bias introduced in AI by the companies that design its algorithms and even if bias doesn´t exist it would be diificult to imagine to have the needed ethical capability introduce in the "politician algorithm". For instance, which should be the weight given in the algorithm to the goal of being reelected in comparison with the goal of protection of the general interest? In order to give a real life exmaple, which course of action would a robo-politician choose in an extreme social divisive situation as Brexit?

Maybe the reason the case why we admit robo politicians to be introduced in the political agenda is because we think that bias and a personal interpretation of the general interest already govern the actions of politicians. This would be a more dangerous sitaution that it looks at first sight, because it would mean just another triumph of the so-called far right populisms and those who rock their craddle. Furthermore, the opposite case, the possiblity of creating a robo politician without bias and guided by a perfect interpretation of the general interest would be equally danger, What would be elections for in that case?

As in other cases of AI applications, robo politicians looks as one of those paths that it would be better not to be explored. The seeds of a dystopian society are clearly embedded in the idea. However, we should be sure some one will go that way. 

miércoles, 30 de enero de 2019

"The Entrepreneurial State: Debunking Public vs. Private Sector Myths" - Marina Mazzucato

The Entrepreneurial State: Debunking Public vs. Private Sector Myths

My rating: 4 of 5 stars

A must read book if you have a left-wing vision of the mission of the State. And if you don´t have it and you are open to new ideas, perhaps it would make you think about your own convictions. Because this book, as its subtitle said, is oriented to debunk some myths about the role that the public and private sector has historically played in making fly innovative ideas. Through a compilation of examples mainly from the digital economy, Marina Mazzucato put some prejudices upside down, showing that public sector has been behind of the inception of the more profitable innovations around us.

We can call the author´s theories as post-keynesians, as she does not only reivindicate the role of the State in economy as the one that makes counter-cyclic investments. The State should have also a role as the sponsor of those projects that the private sector does not dare to tackle due to its high initial investments. The author shows that this have been the case of renewable energy devices or the Internet (both as a network and some of the most succesful applications such as the WWW or the iPhones).

But the book tackle also one of the topics more cheerished by the letf-wing policies: the need of a more fair distribution of benefits. In spite of the economic success of the innovations mentioned above, few of this profits has returned to the State for promoting new innovations (some succesful and some not) or providing services to the citizens, particularly to those affected by the disruptions caused by the innovations. The author points to this failure of the innovation system and calls for the development of mechanisms to solve the problems (including some examples).

However, some times you may feel that the book is longer that needed. I also have that feeling. But thinking over it twice, I think that the lenght is fully justified: Without plenty details it would be an impossible mission to debunk the myths. The book is written not only for those convinced of the need of an entreperneurial state, but also for those who believe in the prevalence of the private sector over the public sector.

To sum up the author thesis that I fully share, "if governments are willing to take the big risks that business will not take, they are bound to fail sometimes and suceed others. But if they do not do it, they will not succeed at all". State has already playes this role and should continue that way. However, we should solve that "risk taking has been a collective endeavour while the returns has been much less colectively distributed".

View all my reviews

miércoles, 23 de enero de 2019

Forget your hopes, work is not going to disappear

Fear to automatisation is in the air. People are scared of losing their jobs due to increasing use of Artificial Intelligence, particularly with the integration of intelligent robots in the working place. According with the Eurobarometer on "Attitudes towards the impact of digitisation and automation on daily life",  74% of respondents expect that due to the use of robots and artificial intelligence more jobs will disappear than new jobs will be created and 72% of respondents believe robots steal peoples' jobs.  

This fears are mainly based in an Oxford University study wrote in 2013 that estimates a 47% of American jobs are at risk of automation, which methodology has been applied in different countries with similar results. Less known is the explanation of the study written by the authors in 2018, which among other things highlighted that this estimation does not mean an employment apocalypse, but only a hint of the scale of the changes that we are facing and the need to craft the appropriate response. As a matter of fact, something similar have happened since the beginning of the XX century with the workforce employed in agriculture, that has decreased from 40% of the population to 2%.

Evidences against the employment apocalypse scenario are beginning to emerge. In spite of its short base of respondents, there are interesting the conclusions of a recent study among American companies that have already incorporated AI as an enterprise tool. Far from diminishing the workforce employed, 40% have increased the number of employees and 34% have neither experienced a decrease nor an increase.

The above mentioned study is nothing more than the confirmation of a warning against neoluditism made by Gartner last year: The apocalyptic predictions don´t take into account how the technology and humans can work together to create new employment. Furthermore, the study also shows a that most companies had move from awareness and early-stage adoption of AI to implementation and determination of what value the technology provides, which also should have influence in a wise usage of AI beyond the mere substitution of humans by machines.

So it would be better to forget for good or bad about the end of work, it is not going to exist. Work is going to change, in the same manner that has evolved since the dawn of humanity, but it will continue to be our daily damnation
palyginti kainas