The dangers of artificial stupidity
This post is also available in: Spanish
Common misconceptions about artificial intelligence
“Artificial intelligence threatens the existence of our civilization“, “Artificial intelligence could mean the end of the human race“, “The AI will be able to write essays at a high school level by 2026 “.
It could be a good start for the trailer of a summer movie, if it weren’t for the words of Elon Musk, the Silicon Valley billionaire who is trying to save humanity, or Stephen Hawking, the most media-driven astrophysicist of our time. The third sentence relaxes the tension a little more, after all, don’t you already copy the work in the institutes from the Internet? Is it really so dangerous that they copy? The same study (from the prestigious Cornell University) says that artificial intelligence will be able to drive a truck in 2027 and write a best-seller in 2049. With regard to best-sellers, I suspect some of them may be written by algorithms.
The strange thing is that people smarter than me persist in a message of fear, in a message that fills headlines and throws us into the hands of other individuals who play along with them, talk about regulating robots with the four laws of Asimov’s robotics, a nonsense that would be funny if they didn’t say it too often. What do you want? Hawking and Musk should leave science fiction to authors like William Gibson, who in 1984 wrote “Neuromante”, a book about Artificial Intelligence in the fight against human beings. Best seller and Time cover, but the eighties are over and Cyberpunk, virtual reality and robots are still far from being dangerous. They are very juicy headlines that can damage the reputation of computer science that works when artificial intelligence really helps humans.
Fortunately, an artificial intelligence expert such as Peter Bentley, disassembles these headlines with a forceful sentence: how are we going to create an intelligence more capable than human intelligence if we fail to understand even the basic principles of the human mind, how are programs that only solve problems for which they are designed to acquire self-awareness? It’s like asking a science calculator to get the dog out.
Much more forceful, Yann Lecun, head of Facebook’s Artificial Intelligence research department, said a few days ago that “current artificial intelligence does not surpass that of a rat. Tough words for someone who makes a living with it.
In 2024, AI may be able to translate texts better than humans, as Cornell’s study says, but the truth is that nowadays, the only damage Google translate does is to our eyes when we read some instruction leaflets. The bursting of the car, internet and soap operas left many people out of work, but it was not the end of the world, but a new opportunity to reinvent themselves; they say that robots and AIs are going to make large sections of the population unemployed. The same thing happened when they invented television or the phone. Artificial intelligence is not our enemy, but our ally in an increasingly complex world.
In almost fifty years of research, the closest a synthetic intelligence has come to mastering man has been in winning Go, an Asian game with more mathematical complexity than chess. The bad darknet languages say that there is code on the net that allows you to make money playing online poker.
Metropolitan London police have invested a lot of money in their surveillance camera network and face detection programmes only work well in very specific situations, so this technology is not yet too useful for finding the needle in the haystack. Almost by chance, they realized that one in a hundred agents – of those poorly paid and questionable eating habits – were able to spot faces thanks to an innate ability to remember faces. When they put several of their superpowered agents in a room to work together, they discovered that they could easily find their suspects in that sea of faces. In any case, I’m sure that most people who spend their days glued to a screen with such an incredibly boring job would prefer their work to be done by machines, smart or not.
Don’t be afraid of what your assistant’s voice on your mobile suggests, but be wary of humans who write biased headlines in newspapers and those who without writing science fiction predict the end of humanity with precision: something they will want, their money or someone else’s money for a good investment. For now, artificial intelligence will help you with small tasks. The greatest danger is that you end up putting too much faith in something that doesn’t deserve it.
Self-conscious artificial intelligence does not exist, but they would already be writing articles like this to hide from the greatest danger on the planet: us.
- “La inteligencia artificial podría significar el fin de la raza humana”
- “La inteligencia artificial amenaza la existencia de nuestra civilización”.
El país, July/2017.
- When Will AI Exceed Human Performance? Evidence from AI Experts
Cornell University; May 2017
- Superior Face Recognition: A Very Special Super Power
- “A inteligente terminator”
- “La inteligencia artificial no va adquirir consciencia de la noche a la mañana”, Peter Bentley.
El país, Julio/2017