Artificial Intelligence or Stupidity? – 08/03/2021 – Suzana Herculano-Houzel

I know “artificial intelligence” still sounds like science fiction, which is about androids and machines that will make decisions instead of humans in the distant future. This mental picture has to be very useful to the interested parties as I don’t see much effort in undoing it when the truth is that other than the human-like robot part this is unnecessary but very useful to the public too employ Looking at the hand, artificial intelligence has been around for a long time.

I don’t think it’s bad, on the contrary. If intelligence is the cognitive flexibility that, among other things, enables us to identify and choose the paths that will keep more doors open in the future, algorithms that help us with this or automatically offer the best options are very useful.

Browsers, for example, are legitimate intelligent route calculators: they consider start and end points, alternative routes, user priorities (faster? No roads? Fewer traffic lights?) And return the optimal choice while using your own neurons to solve other problems.

In this category of algorithms promoting good alternatives, researchers at Uber’s Artificial Intelligence Laboratory recently found that adding two simple principles to their unsupervised learning programs makes all the difference. In the article published in the journal Nature in late February, the team reports how maintaining previous partial solutions and exploring new possibilities from a return to those partial solutions can be incorporated into the algorithms to make the benefit programs better than humans. in games from the Atari universe, which have become one of the success milestones for such learning algorithms.

And when you learn what it takes to create really smart programs, that is, to explore and maximize opportunities, we also learn what is important to maximizing our own success – in this case using what has been referred to in the past as new works has starting points for exploration.

This is completely different from the artificial stupidity that has spread through social media and other apps that “learn” what the user then likes to offer … more of that. The possibilities offered by the networked world are too extraordinary for a user to be able to cloud them with programs that only limit and restrict their horizons and perspectives.

PRESENT LINK: Did you like this column? The subscriber can release five free hits from each link per day. Just click the blue F below.

Leave a Reply

Your email address will not be published. Required fields are marked *