Language lessons across the UK & Ireland

Call us! 0203 650 19 50 / +353 (0) 1 440 3978

Is A.I. En Route to Mind Manipulation?

Be honest. If it wasn’t for the fabulous Benedict Cumberbatch, it is likely that you would neither know anything about the enigma machine, nor would you know about the genius that was Alan Turing.

The Imitation Game, a movie based loosely on the events of Alan Turing’s life and work, was actually the title of a test performed in early studies that eventually led to the Turing Test. And the Turing Test, the ultimate test of man against machine, is arguably the first step in demonstrating true machine intelligence. The question is, who is manipulating who: us our devices, or our devices us?

Photo via Flickr

Introducing Eugene…

Because what better way to delve into machine manipulation than to talk about the algorithm that passed the Turing Test. The Turing Test is essentially a way to determine whether or not a computer is intelligent enough to fool a human into thinking it is another human. Strictly speaking there are a series of Turing Tests with increasing complexity, but the reason Eugene made the headlines when he passed the test was because he was the first since Turing proposed the idea back in 1950.

Eugene Goostman is an algorithm chatbot that presents as a thirteen-year-old Ukrainian boy, which is in itself a genius tactic. Because what thirteen-year-old can possibly know everything and wouldn’t be a little awkward speaking to a strange adult over the internet - especially speaking in his non-native language?

The test pit Eugene and other human interlocutors against a panel of judges who had five minutes to determine whether their conversing partner was human or machine. Eugene convinced 33% of the judges that he was indeed, human, and therefore the threshold of 30% was passed and the initial stages of the Turing Tests beat.

Predicting the future…

Photo via Flickr

No one is immune to the pain of an auto-correct fail. There are entire websites dedicated to such delights, leaving readers weeping in hilarity and sending in their own examples. However, as you have probably already figured out for yourself by now, these errors our devices soon learn and adapt to so that we can happily write certain words in our messages without it being chastely being changed to ducking (you know what word we mean).

Google has for the longest time been king of prediction when it comes to knowing what we want to write. Not only does the search engine suggest the very word we are typing, it also predicts what we are planning on writing next. Its Microsoft Word equivalent Google Docs has also recently taken on a feature called Explore which takes that assumption one step further by suggesting words, images and articles that may be helpful to you based purely on the words you have already written.

Facebook, ever innovative in its social media expansion, launched an AI tool this summer called fastText, which in simple terms interprets and ‘reads’ large blocks of texts. What does this actually mean? Well, if you are a Facebook user (because we don’t wish to presume), you may have noticed that your Timeline has of late become more you friendly. You probably haven’t seen NSFW content that you haven’t willingly searched for yourself, nor any spam in a while. You are also more likely to have seen better suggestions for things you might want to look at based on those that you already do.

Into the deep (freezer)...

Photo via Flickr

Smart refrigerators have been with us for a little while now, giving information like weather and local news on a touch pad attached to the front of the fridge that can also double as an interactive note board system for the family. This year we saw this useful ‘tool’ go one step further.

When its latest Galaxy model isn’t exploding on our coffee tables, Samsung is on par with the technology giants we have just mentioned. Back in January it launched a refrigerator that tracks the contents of your fridge and orders you replacement items when you need them. Now, this might seem like something off the USS Enterprise, and yet it is not quite that interpretive just yet. Products are ordered from one particular source online, so it isn’t as though you are sending out a little robot helper down to the local butchers and markets to pick up your fresh organic goods. It is, however, progress.

Smart heating allows us to control when our radiators come on and go off through our smartphones, and Sky channels can be programmed to be recorded with a simple few clicks on our touchscreens. Technology is slowly creeping in and taking over aspects of our lives we didn’t necessarily know we wanted taking off our hands but now that we have them, we probably don’t want to do without.

Learning a new language? Check out our free placement test to see how your level measures up!

Hearing out Hawking…

Everyone’s favourite astrophysicist Stephen Hawking is one of our smartest cookies, so when he talks of his fear, or prediction that technology may go full Skynet on us and one day supersede humanity, perhaps we should listen. Sure, it’s great to have all of these tools to hand that prompt us to buy things and takes the very words from our mouths so we don’t even have to think when we type. But what happens when these helpful suggestions become a little more targeted, or, dare we say it, manipulative?

Take for example recent accusations that Google is manipulating its searches in favour of Hillary Clinton in the current run up to the elections. By slowly changing the things we search for and suggestions of where we should shop, what is to stop AI-savvy companies from producing algorithms in favour of their products? What is to stop entire nations doing the same with our search engines to promote their political views? And what happens when the machines themselves start thinking for themselves, manipulating our own way of thinking?

We will leave you with that sobering thought.

Next time on our peek into AI advances we will be looking at the potential obsolescence of the humble computer programmer and computer programming languages themselves. Until then...