Nothing is free with AI

Nothing is free with AI
Nothing is free with AI

The training of Artificial Intelligence advances so quickly that not even the government agencies in charge of its supervision are capable of reacting in time. Thus, the privacy of our data and even private businesses for which there is still no regulation is at risk, remember the following column for CIPER: «For the SIA to work, it must have a large volume of information from all dimensions of life. human. […] That means that, in some way, we are all useful for this purpose; especially, the most gullible.

hUntil next June 26, there is a deadline for each user to tell Meta (ex Facebook + Instagram) not to use data from their personal social media accounts. to train Artificial Intelligence (AI). This is a relevant issue in terms of privacy, but which you probably weren’t aware of or, if you did, you don’t see a big problem with it. AI is trained with language patterns, data and information from any region of the world, and Latin America—particularly the coastal territories, with their idioms and slang in the language—will be no exception. We are facing the transition from a generic generative AI to a generative AI personalized that, in the words of writer Yuval Noah Harari«hack» our own language and thinking, creating predictable patterns for uses we do not yet know.

We remember a song by the Cuarteto de Nos and this realistic adage: “If the alms are large, even the saint distrusts.” As we use powerful AI applications every day, let us remember that “nothing is free… nothing is free in life.” We are paying with something.

We can say that AI had successive phases of maturation until reaching the current moment. Based on its level of development today, it is difficult to make a prospective judgment for the near future, since advances in the matter are exponential. The classification of AI today focuses on the development of Narrow Artificial Intelligence (NAI), the pursuit of General Artificial Intelligence (GAI), and dream of Super Artificial Intelligence (SIA). IAE is called this because of the limited field of its action, focused only on a human task (for this reason it is often called “weak AI”). In the IAG, the cognitive conditions of the human being are equalized and with it any human activity can be developed (“strong AI”). While the aforementioned AIs focus on one (weak) or all (strong) human tasks, the SIA surpasses them; that is, it goes beyond human intelligence. For philosopher Nick Bostrom, the SIA is “an intellect that is much smarter than the best human brains in virtually all fields, including scientific creativity, general wisdom, and social skills.” There are constant changes in the apps, programs and services that use it that may seem imperceptible, but are enormous due to the specificities of their transformation.

When the great wave of dizzying advances in AI came in 2023, the data and information with which it was fed were facts, phenomena and people from developed countries, which generated biases in their outputs or exits. For example, there were errors in generating images that did not fit with the breed or phenotypic characteristics of people from Latin America, and then we saw white people carrying out managerial activities and African Americans with a tendency to commit crimes. The world demanded an AI free of racial or xenophobic prejudices, so that it could see itself represented in all its diversity. Although voices were raised from the academic, scientific and political spheres, they were not enough to stop the mighty river of updates and new developments that continued until today, in a wild competition to capture the attention of all users, of all ages. and geographic location.

Today, the development of secret (or very secretive) projects continues at full speed, for which it is necessary to fill gaps in processes and information. For the SIA to work, it must have a large volume of information from all dimensions of human life. You have probably already started a human thought cloning megaprojecttrying to emulate it and translate it into powerful chips processing large volumes of data.

That means, In some way, we are all useful for this purpose; especially the most gullible. To capture biometric data, they will ask you to scan your iris [imagen superior]and maybe it will be profitable for you, in exchange for cryptocurrencies that you can convert to euros. European government entities They have initiated ex officio investigations in this regard, but the accelerated pace, the geographical migration of these companies and The rapid transformation in the means of acquiring personal data is advancing so rapidly that States are left incompetent in the matter.. In addition to the above, the legal departments and unlimited resources of technology companies make legal battles long (not to mention fruitless).

Harari has said it: «We have invented something that takes away our power. And it’s happening so fast that most people don’t even understand what’s going on. “We have to make sure that AI makes good decisions about our lives.”

There are people who still believe that AI cannot reach the language of the territories in the world; for example, from Latin America and its great native linguistic wealth. Even within Latin America itself, each country and subnational territory manages its own cultural expressions that would be very difficult for AI to replicate. Well, they are not visualizing the transition from an AI trained with general data and information to one that is currently being trained with particular data and information, perfectly capable of capturing our idioms and even our humor until it seems trained in a neighborhood in Mexico City. Mexican, from Medellín or Valparaíso.

For Latest Updates Follow us on Google News


PREV A cortado with a bitter taste | A player from the Spanish national team launched his specialty coffee with a logo similar to that of the Nazi SS
NEXT Bogotá is among the 15 cities with the highest cost of living in Latin America