What Happened to Facebook’s AI?

Facebook was forced to shut down an artificial intelligence engine (known as AI) after discovering AI had created its own language. In 2017, Facebook caught two chatbots, Bob and Alice, speaking their own language to each other.  Many scientists have stated that AI could lead to horrible and violent consequences. Even Stephen Hawking cautioned in 2014 that “AI could mean the end of the human race” (Forbes).

The consequences that AI could create are severe and could be extremely dangerous. Our entire world is connected and especially in the future, almost everything will be connected. The fact that an AI would be smart enough to figure out we are all connected and will use that to their advantage. Computers can distract an abundance of data from a user at a rate a lot faster than humans will ever be able to. The possibility of a machine taking over is unfortunately not unrealistic.

Artificial intelligence can help improve human’s lives in a great degree, but it is hard to tell how far to let them go. They have the potential to make our lives significantly easier, as well as the potential to ruin mankind.


9 thoughts on “What Happened to Facebook’s AI?

  1. Your article first caught my attention because the title mentioned Facebook. Facebook is a site that I frequent on a daily basis so I am rather surprised to have not known that Facebook tried to connect with AI. Hearing about robots or machines taking over the world is something I’ve only read in novels or seen in movies, but reading this article reminded me that this is in the definite realm of possibility for the future. I am a firm believer of utilizing artificial intelligence to our advantage. Artificial intelligence can be beneficial. For example, AI improves data management, cyber security, and the internet of things.

    On the other hand, I find it extremely alarming that AI was creating its own language that is incomprehensive to humans. If AI was not programmed to create a language, how was it able to do so on its own accord? Since technology has become that advanced, it is our duty to create boundaries that will limit such things from reoccurring. Without stepping in, how much further could AI develop without human assistance?

    According to a government commissioned report, AI could add 630 billion pounds (the equivalent of $837 billion dollars) to the U.K. economy by the year 2035. The report detailed that their goal is to advocate for students to pursue a degree in the field of artificial intelligence. Additionally, the report recommended for people to take conversion course in order to acquire skills in machine learning and better understand the technology that is being dealt with.

    If artificial intelligence is being positively promoted on a global scale, I wonder where this will lead our country and the rest of the world in the future. People are going to have to learn more information and become open of artificial intelligence before fully being accustomed to potential new changes. This would be a big step because I believe people find artificial intelligence to be a controversial topic.


  2. What I find super scary about this is that I’m a frequent user of Facebook and had absolutely no idea about this. So, while I was on Facebook posting all of my pictures and sharing different things, this creepy “Al” was creating its own language. While I don’t understand exactly what this means, I can’t help but wonder… was my information somehow used to create this uncontrollable language? Or, worse, what if Al was able to reach top secret information, or perhaps someone’s posts about creating bombs? What if Al got his inspiration from evil? This post really highlights how powerful IST is. While it scares me to hear about, it’s also so cool that I’m entering a field where I can get a look behind the scenes of this craziness.I found a Forbes article that I definitely agree with, discussing how we need to take a few steps back and let humans take control back into their own hands. But… is it too late?


  3. While artificial intelligence does have the potential to make our lives significantly easier, I feel like there needs to be a limit. What will humans do if machines are capable of doing all of their jobs? In a study conducted from Oxford, Yale, and AI impacts, they surveyed 352 machine learning experts to forecast AI progress. This study showed that there is a 50% chance AI will outperform humans in all tasks in 45 years, and also a 50% chance that machines will take all jobs in 120 years. While it is alarming that two chatbots were able to create their own language and speak to one another, I feel that the more realistic fear is robots taking over our job industry. Industrialization and modern technology has made our lives much easier, but do we really need to get to the point where they do absolutely everything? 120 years is not a long time to possibly have all jobs conducted by robots, and humans will become essentially useless.

  4. This article caught my eye because I have heard about the possibility of artificial intelligence eventually taking over the world. I never knew it was a reality and I kind of brushed it off as a hoax, but as I read more articles on the topic I found that it is more of a reality than I expected. Newsweek wrote an article about the AI takeover and it had several interesting facts listed within. Newsweek stated that in the next 45 years there is a 50% chance that AI will be able to preform all human tasks better than actual humans. This is eye opening because I will hopefully still be alive when this comes around and it could have a drastic impact on my future. This also poses a huge threat to the new generations as finding jobs that can’t be done by computers may be rare to come across. The article also stated that within the next 120 years all human jobs will be automated. I am sure AI has many beneficial things to bring to the table, but we may want to start looking at the negative sides to this extremely advanced technology.


  5. I looked at other companies AI’s and was interested to see what all they can do an saw that Wells Fargo’s said to sell popular stocks like facebook and google (which is why it came up when I googled facebook AI) I think that’s another way that we may let AI’s lead our life when we shouldn’t. If the stock market becomes almost completely predictable by these machines then it ruins the whole point of the market. The AI’s are getting way to smart and creating their own language is proof of that.


  6. This post was particularly interesting to me and I ended up doing a little research because I was curious to see how exactly the AI chatbots altered the English language to make it “more efficient.” Although Snopes debunked some of the sensationalized articles that were written on this topic, it did highlight the fact that there was no reward built into the programming for only utilizing the English language even though Facebook intended to use this software for AI to human interaction rather than AI to AI negotiations. A curious example of this language alteration can be seen in the quote from Dhruv Batra who is a research scientists from Georgia Tech who is working for Facebook’s AI Research or FAIR. He had the following explanation: “Agents will drift off understandable language and invent codewords for themselves. Like if I say ‘the’ five times, you interpret that to mean I want five copies of this item. This isn’t so different from the way communities of humans create shorthands.” Although the researchers chose to stop this program from speaking with itself, it is interesting to see how this invented communication can mimic human shorthand despite seeming like nonsense to a typical observer. Obviously there are some nuanced aspects of human language that simply cannot be replicated by AI.


  7. This was interesting for me to read because I had never heard of Facebook’s Al! Just this morning, however, I came across an interview on Twitter with a 13 year old boy (genius?) who was hired by Google. He specializes in Artificial Intelligence. When asked by his interviewers what he thought about the possibility of AI becoming smarter than the human race one day, his thoughts were slightly calming. He explained that he refers to that possibility as the “point of singularity”, but due to the fact that Artificial Intelligence is so primitive and new, we may not ever reach that point. If we do reach that point of singularity one day, it is going to be in the extremely distant future.

    This was interesting because immediately after watching this interview, I came across this very post on the IST Blog stating the exact opposite, that Facebook’s ex-AI was already well on its way to gaining its own intelligence beyond human capabilities and had to be dismantled for that exact reason. While Artificial Intelligence can be a really beneficial thing to society and can make things extremely easy for us (i.e. Siri, Google, Watson), we have to constantly be on the watch and we have to consistently be cautious of the fact that they must be kept in check.

  8. While I agree that AI programs, such as Facebook’s, that can adapt to better deceive people, and even create its own language that people cannot understand is a scary reality, I do not believe that it is plausible for it to ruin or take over our world. Facebook’s program was designed to become the best negotiator, and that’s exactly what it did. Although it was not programmed to lie, it discovered this as a means to achieve its goals. But the program was set up to deceive people in the first place, unable to be distinguished from a human being online, and just like human beings, it learned to lie. The truly unsettling result of this research was the fact that the program created its own language to become more efficient, also something that it was not programmed to do. However, once this was discovered the code was altered so that the program only spoke english. That being said, although AI is capable of adapting and learning, it is still entirely under the control of its creators at the end of the day. Therefore, it is not the technology we should be afraid of, but rather who is using and manipulating this technology.

    Information retrieved from: http://www.newsweek.com/2017/08/18/ai-facebook-artificial-intelligence-machine-learning-robots-robotics-646944.html

Leave a Reply