miércoles, 22 de febrero de 2017

We need to talk about #robots

Few weeks after the beginning of 2017 it seems clear that robotisation of the production will be at the centre of many debates in the political arena. We have overcome the stage when we exchange opinions about the myth and reality of a future where robots replace humans for many tasks. Now we know that it is going to happen, the question now is what are we going to do to adapt our society to the new scenario.

We can not aspire to stop the deployment of a technology that promises a a 38% revenues increase and 37% costs drop by 2020. Robotisation is bring also the not so glossy promises of the disappearance of nearly 50% of the current jobs. But the later figure does not necessary means 50% raise in desemployment if the decrease in cost is used to generate new business opportunities around the traditional jobs. We should not think of human replacement but in a augmentation of human capabilities, what the Davos chief call "humanisation over robotisation".

We need to bridle the robot revolution and change the presumed bleak future. The 2017 WEF´s Global Risks Report identify robots as the technology that has "the highest potential for negative consequences, and also the greatest need for better governance". In this scenario the call for an EU regulation done by the European Parliament in its legislative initiative "Civil law rules on robotics" is a step in the right direction. We need to achieve a common understanding about what is an intelligent robot, the taxation on robot labour activities, liabilities in case of accident and even robot rights and the need for a robot personhood

Besides regulating, we need to educate people on what is a robot. It is worrisome that "people are attracted to financial advice from robots because they consider it impartial". We need to extend the understanding that robots are product of human mind and therefore its intellect biased by its creator principles, values and prejudices. Otherwise, it would be difficult to understand the need for diversity in the AI production force.

We are just at the beginning of the debate, but we need to talk seriously about robots. Beyond the technological debate there is a social debate that we should not allow to be skipped.

miércoles, 15 de febrero de 2017

The other faces of digital lilliteracy

The digital transformation of the economy and society is stressing the  foundations of all the countries. Technology has unleashed a subtle earthquake, but with some extremely violent displays from time to time. There is a paradox. Sometimes, the more developed the country, the more violent is the impact of these strong landslides.  This is especially truth when the quake hapens around a weak point that we thought that had been repaired forever  many  time ago. This have been the case of the affair related with the viral spread of fake news during the last US election

Although fake or not accurate news have existed since the dawn of humanity, it was thought that their impact on the developed societies was small now.  The constant evolution of the political systems had consolidated an independent and truthful media system reinforced by the existence of a critical and well-educated population. On one hand, traditional publishers were conscious that publishing fake news was a major risk for their business, as they could lose credibility and could face legal actions. On the other hand, the majority of the public was equipped with the skills to identify the possible bias of the media and to apply a critical view when they receive any piece of news. The balance has been broken with the appearance of the new digital media publishers as Facebook. 

According with a report from the Reuters Institute, there is a growing happiness with having the news served through an automatic process. The consequence is the reinforcement of the trend of having social media as the main source of news. The apparently disappearance of the human factor replaced by algorithms provide the whole process of presenting the news an aura of bigger independency.  This fact shows clearly a misunderstanding of what an algorithm is. As a product of a human mind, the algorithm inherits the bias of its maker.

The blind faith on the independence of algorithms is not reduced to those used on media sector. A recent study published by Accenture claims that bank customers would preferred to be advised on their investments by machines. People attracted to financial advice from robots because they consider it impartial. It is extremely worry that this big misunderstanding about the nature of algorithms is expanding to any area.

There is growing concern for the development of the digital skills among the governments and international institutions. For instance, the aim of increasing digital capabilities was included both in UN 2030 Agenda and the Digital Single Market Strategy. Generally, the success of this kind of actions is measured in the number of Internet users, the amount of digital services users or the number of PCs per person.  What shows the misunderstanding of the nature of algorithms is that digital illiteracy has other faces that we rarely pay attention and has bigger impact in our social life.

miércoles, 8 de febrero de 2017

#Workalipsis : The premier of the movie has been delayed

I don´t know what´s happening around you, but near me everybody is suddenly frightened with robotisation, artificial intelligence and the end of work. I´m too lazy to search when i wrote the first post on the issue, but certainly it was more than  two years ago. I am still thinking that the jury is outside the room and we will have to wait which will be the final balance.

However, more and more articles foreseen a "workalipsis". The last dark prediction I have read is related with driverless cars. The author does a thorough review the things and professions connected strictly or loosing with the fact that a human being drives the cars and the conclusion is that 128 things will dissapear in the future. Among these things there is a huge collection of job posts. 

The bad news are that these predictions are wrongly focused. There is a confusion between the change of a job post and the end of a job post. AI and the digital transformation will augment the human capacities to perform better some jobs and as a result some major tasks in these jobs will become lesser important and replace by others. Returning to the article "128 Things that will disappear in the driverless car era", take the case of bus drivers. The bus driver will stop to have the responsibility to stay behind the wheel but he will have time to provide the passengers other kind of services (e.g in a school bus solving doubts of the lessons to the children, in a touristic bus strenghtening the knowledge of the local customs to the tourists, ...) but this do not necessary means the dissapearance of the job post.

In this video, Tim O'Reilly gives some clues on why we will not run out of jobs. To begin with, we will still have to think on new ways to solve human problems or to repair the mess we make on a daily basis (e.g. wars or climate change). Coming to our closer environment, we will have to think on how to transform our analogue goods and services in digital goods and services. This is the first of the 5 tips for digital transformation given also by O´Reilly:
1. Build a digital organization, not just digital products.
2. Keep your eyes on the road, not just on the map.
3. Never stop learning.
4. Remember that the future starts today.
5. Don't replace people. Augment them.

These 5 tips reinforce also the idea that we are not facing the disappearance of jobs or our society, but a change. We may think that digital transformation is a deeper change than previous changes, but that it is because we are personally invlove in it. Try to think as those who lived the appearance of the steam machine  or the electricity. For the worst or the best, workalipsis will not come soon to your town.

miércoles, 1 de febrero de 2017

Artificial Intelligence: A story of police and thieves

There is an unstoppable move towards automatisation. Besides the routinary jobs that do not need special skills to be performed, news show us that even the more sophisticated jobs are not safe of not being taken by robots and algorithms. On one hand, the role of middle managers looks to be redundant soon in hedge funds firms where the every day decisions about investments decide the future of thousands and millions of jobs and people. On the other hand, "robolawyers" show their hability defending people in court in a more efficient manner than humans taking into consideration a huge pile of previous cases.

As times goes by, it looks that the expanding role of artificial intelligence is something more than the wish of geeks. IBM, Google and many others are getting ready for a near future where nearly every decision has something to do with AI. The global artificial intelligence market was valued at US$ 126.24 Bn in 2015 and is forecast to grow at a CAGR of 36.1% from 2016 to 2024 to reach a value of US$ 3,061.35 Bn in 2024. However, it should be highlighted that it will be difficult to confirm this predictions as AI it is more a feature than a product.

The case for AI is more than justified in massive services. Take the case of Facebook and its obligation to fight fake news that recently has been demanded by the roar of the society. The task to clean the facebook feed from fake news only in Germany requires daily 600 people to perform 2000 deletions. But the more massive the service the more critical is to eliminate any bias in automatisation, especially in the area related to freedom of expression. 

The more fields find an application to Artificial Intelligence the more it is needed to embed ethics in the role of algorithms. Furthermore, this brings a new paradox. If AI is pervasive in any process and decision, and it looks it will happen, the role of watching the ethics of algorithm ought to be encommended to other algorithms. As it was expected, some scientist are working on this issue also.

So algorithms and artificial intelligence looks at the same time as the problem and the solution. But better not to start with the question who oversees the guardian in the artificial intelligence world.
palyginti kainas