miércoles, 27 de marzo de 2019

Expanding the debate: Digital Ethics

According with the Cambridge dictionary, "ethics" means set of beliefs about what is morally right and wrong. Therefore, the term "AI Ethics" means the set of beliefs about what is morally right or wrong to do using AI technology. "AI ethics" has become one of the central elements of the debate on AI development and deployment, at least in Europe, Between the US entrepreunerial vision and the Chinese state-leading conception of AI, an ethical AI has emerged as the cornerstone for the European vision.

Certainly, it should be a central debate of AI development which are the non-technical limits on all the elements of the value chain, from the design of products to its usage without forgetting its construction. However, the debate is incomplete. Why we should ask ourselves only about "AI ethics" and not open a broader debate on "digital ethics"? Computers have been around us since the 70´s of the last century and massively since the year 2000, so perhaps it is high time to open the debate on "digital ethics". Even more, perhaps the root for the global mistrust and neoluddism is the absent of this debate in the last years.

Ethical questions and issues do not raise only from the usage of AI. The massive development of disruptive technologies poses question to us at every moment if we stop to look around.  Is it within the ethic limits the commersialisation of DIY-neurohacking devices without a knowledge of the long term effects over the brain? And what about the massive deployment of IoT based sensors around the city and capturing citizens generated data without their previous knowledge and consent? Or have we thought in of the implications in third-world countries of the constant renewal of en user devices without imposing circular schemes of manufacturing? Only three examples of dilemmas, and we have not included yet any related with the digital services we daily use.

Another unbalance of the articles and papers of the debate on AI ethics is its usual focus on the design of products and services. It is obvious that the deployment digital ethics should start by making it a boardroom issue. We would not have our democracies at risk now without an executive that had decided many years ago that digital targeting marketing was a good product to offer to anyone, included those with obscure political intentions. But the responsibility on digital ethics is spread across the whole digital value chain.

Sometimes, when we think in digital services as Facebook or Google we tend to personalise them end-to-end in their CEOs. Few times we think that these and all the products could not exist without the existance of engineers who has worked in their construction or marketing specialists who have design clever advertisement campaigns. The universalisation of digital ethics are also their responsibility. Microsoft workers provided an example to follow in this field with the letters they wrote to the executive staff signalling their unhappiness with some contracts won by the company to develop applications for the Army and the Immigration Control Agency. The message of the workers was clear "we refuse to create technology for warfare and oppression". Perhaps someone should think of including ethics in the engineering curricula.

Last but least, we consumers have also our share of responsibility on digital ethics. Maybe sometimes we have as an excuse not knowing all the implications of choosing one product or service, but even sometimes we fully understand them and we still use services that are based on precarious work due to its cheapness. Nevertheless, the starting point for demanding consumers more responsibility on digital ethics is a requirement for better and clearer information on products anda services to their manufacturers.

Digital ethics should neither be limited to one technology nor be the responsibility only of the boardrooms of tech companies. A wider debate should be opened and all the actors be provided with the right tools to be aware of their share of responsibility.


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.
palyginti kainas