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. 
palyginti kainas