Who Are Going To Be The Robots Teachers?
I read an article not long ago, Where Should Machines Go To Learn, which got me thinking about automation, and artificial intelligence.
I grew up under the shadow of Skynet. Judgement day, when Skynet became self-aware and decided to eliminate the human race, either August 29 1997 or July 29 2004, depending on which canon you want to accept as being true. As a young boy I always preferred the Stormtroopers to the Jedi, and I’d suggest that there are boys out there now who want to be the Terminator, not John Connor — Kylo Ren, not Finn. Perhaps there is a breed of young boys who just want to be on the winning side.
Automation and AI are going to disrupt the world like Henry Ford disrupted the industrial world in 1903. Maybe not to the extent of extinguishing the human race, but nonetheless. People believe that automation means jobs for people will disappear. Every task that is handed over to robotics means one more person out of work. I don’t believe that, however, because I choose to trust in human resilience, and in our ability to find the alternative.
As an example, Automation and robotics doesn’t work if there is no one in the back office to train the AI, or to develop the content so the AI can become more intelligent. The article I read was asking where this content would be kept, but in my opinion a more important discussion is around who we trust to keep it.
Who’s Going to be the Teacher?
The question of who will teach the AI raises the important consideration of legitimacy. I’m not sure I’d be very keen to have a conversation with a robot that has learnt to speak by absorbing YouTube comments and the rantings of Facebook trolls. So who decides on the source of the information with which the robot will teach itself?
Acquiring knowledge however, is not just about finding a suitable source. The process of acquisition is also important. It’s because we learn in a certain way they we are able to make good connections, to see the bigger picture and make good decisions.
So who will define the process of learning in the Automated future? Who will define the Corpus of Knowledge?
Who’s Going to be the Guardians of the Library?
In the dark ages (before 4G and Wifi), scientists and scholars maintained the knowledge of humankind, storing everything in massive libraries built within impenetrable fortresses (I don’t know how true that is, but I’ve seen The Name of the Rose and it feels pretty true). They were vigilant in their preservation of this knowledge base, and were careful not to share too much. They believed their intent was to protect society, as they were convinced (in their intellectual arrogance) that the crowd was not intelligent enough to understand the scope of everything.
Today we are in a world governed by massive knowledge collaborations. Wikipedia is an incredible attempt to create a large-scale, shared library by capitalising on communal knowledge. It’s the new Alexandrie Library. However, if you look closely at the stats you’ll see that the encyclopedia is managed by only a very small set of people. But what is their legitimacy in producing this content? Who chose this particular group of people to be to protectors of our knowledge library? Have we finally found our way back to the dark ages?
And even if we have not, and even if the moderators of Wikipedia are qualified and generous with the content they oversee, are we sure this content is objective? Are we sure the truth found there is really the truth.
The New Laws of Automation and Robotics
Years ago experts all over the world followed Isaac Asimov’s lead and recommended a charter of ethics be applied to robots and robotics. It’s true, the charter was intended to avoid the misuse of automation, rather than protect us from robot overlords, but still…
Since then, however, nothing has really been done to implement that — there are no rules yet in place about the transmission of knowledge to robots. So yes, it is still pretty cool that you can have a lengthy conversation with the Snapchat bot, and you can even try insulting it to see of the configuration is detailed. But the deeper ethical questions about artificial intelligence are still pending, and this is something that needs to be addressed in the next few years.
Unfortunately, this article isn’t full of answers, but instead loaded with questions. That’s because the answers aren’t clear just yet. This is no doubt a topic we will return to before too long because I believe there is a lot of development that is going to happen in this area over the next few years.
