• The Final Interview - Read All Encounters
  • Jia Tolentino. "Being a mother in these times is an act of blind optimism"
  • Eva Amaral. "We are experiencing a bombardment of sadness"

Inma Martínez (1964, Valencia). Artificial Intelligence expert and visiting professor at Imperial College London, since 2010 advising the British Government and the European Parliament on this matter.

Can Artificial Intelligence (AI) and Big Data Help Curb the Coronovirus Pandemic? Sure. And not only AI, but also other digital technologies. But above these are the data. In the data is the pitched battle that is going to decide who is going to gain ground to the virus and who is not: who has real-time, updated, IA-ready data (that is, in formats ready to be analyzed), and above all, verified, it will be the one that manages to extract the informative and strategic level of the analysis. Can you give me an example? Spain registers mortality rates well above other countries. And not because we have a worse health system, but because our contagion count is much lower than what actually exists, so the percentage skyrockets. This is the true interpretation, and I am dedicating myself these days with teams of independent scientists in Europe, answering the call from governments and universities to propose new hypotheses and methodologies to the extraction and accumulation of data. The really valuable data is not the deaths or how many infected we have per day, which is data that comes to us with noise, delayed and through unreliable methodologies, but data on the level of contagion, the rate of spread of a person's virus others. There are already machines programmed to decide on life and death. In this pandemic, and in the face of the collapse of the ICUs, should the decision of who to attend to be left to the algorithms? Let it be clear: the algorithms are programmed by humans. If I train an algorithm to determine which people should be allowed to die, I have to give them operating parameters that I determine, and not the program itself. In the United Kingdom, the health system is already executing decisions like this based on a system of awarding points to each patient according to their ailments - high blood pressure, diabetes, mental health ... - and tabulated by medical personnel (and not by an AI) and patients with low scores are admitted as a priority in the ICU. This will leave many over 65s out of intensive care. This system has not been developed by an AI system because it is not sophisticated, but I insist, would it be better or worse to leave the decision of who to attend to the AI? If we had to trust who is better equipped to determine this type of decisions, without a doubt, an AI system would have much greater capabilities than a human being to calibrate dozens of variables to consider. And depending on how and with what kind of ethical rules this intelligent system is programmed, your answer to the question "who do we leave out of the ICU?" It could be: "Nobody". Can a machine have a higher ethical sense than a human? A machine sometimes has better availability for ethics than a human whose calculation powers prevent him from considering thousands of other variables that would help to elaborate an adequate strategy. be used to detect the advance of the coronavirus early? This is precisely what is wanted with the app that some governments, universities and technology giants are developing taking as a model the app that the Singapore government has used to prevent infections in that country . Being a mobile application, the algorithmic models that are being proposed are typical "spread" models of social networks. This virus reproduces through social contact, and nobody better than engineers who have created social networks to train algorithms for this type of prediction. Could AI be used to forecast the true extent of the epidemic in Spain? Unpublished data but I am aware that the government has - because mobile operators control in infinite detail the movements of people and some have passed that information on to the governments- have helped predict the rate of contagion based on how it has been occurring so far. We train algorithms to predict by feeding them historical data, and teaching them how previous events have occurred, with all the wealth of variables we can add. We know, for example, that the latest infections are between members of the same family and not by third parties, because population movement has been prohibited. Sadly, too, many have occurred among health personnel in the course of their work. Non-essential work has been resumed. Does the AI ​​have anything to say? The Government assumes that the population that has returned to these jobs is exempt from the virus because they have spent 15 days at home, but it is counterproductive to allow the population of any age to enter work knowing that in 90% of cases, patients older than 60 years have a greater chance of complications and death. Variables such as province and community and, above all, the age of people should be taken into account. The main problem for the use of AI in the face of this epidemic is that it poses serious problems of violation of personal rights and freedoms. Can you avoid it? The problem with developing an app that establishes the proximity of individuals is that it lends itself to governments skipping many of the clauses protecting personal data. The data must be anonymous: for example, what matters is how close you are to others and not if you are in Mercadona or Carrefour, that only matters to those in marketing and not to a doctor who will save your life , I explain? In Spain, apps in circulation require the collection of DNI data, and this will put many people back. And for this app to really work, it is calculated that 60% of the population has to download it, enable it with a Bluetooth connection, let it collect data associated with the proximity of other people since we anonymously declare if we are carriers of the virus to be able to inform all those with whom we have had contact in previous days. If great concessions are made in the field of rights and freedoms in the name of the fight against the Covid-19, then will it be difficult to back down? convince otherwise. But we are fighting so that many companies already begin to back down.

According to the criteria of The Trust Project

Know more

  • Final interview
  • Covid 19
  • Artificial intelligence

Technology Google publishes statistics on the effects of confinement in each country

Technology Graphene engineering for pulmonary ventilation against Covid-19

Martin VarsavskyThe brain of the 'app' against the Madrid coronavirus: "Allowing the 8-M marches was a very serious mistake"