Facebook is no stranger to controversy, and that’s putting it mildly. Cambridge Analytica would have been enough of a PR nightmare for any company to make huge amends when it comes to your privacy, but it’s seemingly issue after issue with Facebook.
Repeatedly compromising our privacy is the biggest way Facebook has continued to fail us over the years, but don’t forget its inability to curb other problems like fake news. Content moderation is even an area where it’s been using artificial intelligence (AI) algorithms, but it hasn’t worked out very well.
Even so, Facebook is taking a larger step into the world of artificial intelligence and robotics. The company is hoping these robots will lead to breakthroughs in AI and now it’s ready to take on other tech giants working toward the same goal.
Facebook’s face-palms of 2019 (so far)
Facebook has been making a lot of headlines recently, from plans to split the company into two, to Mark Zuckerberg proclaiming the “future is private” after apparently learning the definition of the word “privacy.”
Then, there are the efforts to combat the spread of fake news and other misinformation. Just listen to this recent episode of Komando on Demand, that deals with the inability of Facebook’s AI and content moderators to determine what’s fake, real, satire, inappropriate or legal.
This week Facebook is pulling back the curtain to detail the various work of its robotics division, and it’s heating up to mean serious competition with Apple and Google. Where Facebook is concerned, it’s teaching its robots skills in the same way you would teach a real person.
Facebook’s dive into AI and robotics
It’s no secret Facebook has shown a past interest in robotics work, but it’s never been clear to what extent. Now, for the first time, Facebook is detailing some of the projects at its robotics lab, saying it’s advancing AI by teaching robots to learn.
That means instead of just feeding huge amounts of data into its systems, it’s trying to teach machines to learn on their own through trial and error. According to a Facebook blog, it’s developing robots that can move around and explore their surroundings, as well as pick up objects they encounter along the way.
For instance, it’s basically letting a six-legged robot named Daisy learn to walk without any relevant information or training. The robot’s AI algorithm learns how to move and improves over time, working out balance, location and orientation through various sensors. They’re working with other robots with hands that have built-in touch sensors on their fingertip.
This type of self-supervised learning will lead to AI that can learn more efficiently and potentially much faster, according to the blog. They say if we want machines that can think, plan and reason the way people do, then robots need to learn for themselves. On a side note, I’ve seen quite a few movies that deal with that topic and they usually don’t end well for us humans.
Continuing work in robotics
Google is also getting back into the robotics game, teaching robots how to grab objects and move around by using machine learning. Details surrounding Apple’s work in machine learning and AI is a little more vague, although it did manage to hire a Google executive to head that department late last year.
For Facebook, smarter AI could mean improvement when it comes to policing fake news, hate speech and other misinformation. At least, that would be a noble use anyway but it’s unclear if that’s how it would be used.
As far as robots go, Facebook doesn’t sell them and there’s been no indication it intends to. If it ever changes its mind, it should probably stay away from calling them “Facebots” since that’s a term used for malicious apps.
Personally, I like the sound of “Zuckerbots.” Just think about the convenience of owning your own Zuckerbot. Instead of having to log in and interact with Facebook, you can just hand your data (and maybe your soul) directly to your new robot overlord.
Hooray for the future.