About Facebook's AI
One of the latest trends in technology is AI. The Accenture Technology Vision of 2017 declares AI the new UI.
Companies like Google, Apple, Microsoft, Tesla and of course, Facebook are investing billions of dollars in AI. Personally, I think that a big part of the future of IT will be AI, but not all. AI requires a set of very specific skills, research, knowledge, teamwork and more research to work properly. We are only discovering its full potential, in both benefits and possible risks. The companies mentioned above hire the best scientists and engineers in the world to make this technology happen.
There are countless applications for AI. The ones that I look forward the most to are related to health. But other sectors include legal aid, financial services and education.
The Jaws movie created an irrational fear for sharks. Irrational, I say, because the chances of being bit by a shark are 1.625 in a million.
Why do I mention this? Movies like The Terminator and The Matrix have created an irrational fear for AI, too. Skynet will not happen nor will the Matrix.
If anything, I am more worried about the economic impact that AI will have in our society, it literally gave me goosebumps when I personally heard Joseph E. Stiglitz (Nobel prize in economy) mentioning this in a conference a few months ago. For instance, in the next 10 years, 3.5 million American truck drivers jobs will be at risk because of AI.
And on a more personal matter, I am more worried, as a programmer, that I have not developed the necessary skills that I require for this field (yet).
Some weeks ago there was an argument in Twitter between Zuckerberg and Musk. Some people are on Zuckerberg's side, some others are on Musk's side. I say this real life, not Civil War. These are two geniuses sharing their points of view, that's it. Musk says the future of AI is apocalyptic, while Zuckerberg has a more positive point of view.
The truth is that the future is uncertain. And I do agree with Zuckerberg on this, creating an irrational fear today is irresponsible.
A few days after the argument, Facebook shut down an internal AI project. The AI was created to understand and reply in English. Unexpectedly, the AI found more intuitive to reply with a pattern instead of English (or so we think). This could be a programming error, gibberish, or indeed, the AI found english stupid and decided to reply with a pattern (that used words in English anyway).
The Internet was freaked out by this, like always. If Skynet was to happen, it is not by a program saying Balls have zero to me to me to me to me to me to me to me to me to. Reality is, Facebook did not panic, and you should not, either.