Sunday 26 June 2011

Dust thou art, but unto silicon dost thou aspire

It used to be the case that when you wanted to use the Internet to find an answer to a query, you had to pick the keywords from the question (much in the same way students are told to approach examination questions) and then Google them. I realise as I chose ‘Google’ as a verb that, at the time, users had to ‘Yahoo’ or ‘Alta Vista’ the keywords as these were pre-Google days; maybe if/when Google is supplanted as the dominant search engine its name will remain as a verb, just as ‘hoover’ has done. The user would then have to wade through pages of results containing the keywords to find the combination which answered their original question. Google’s search technology has, without doubt, improved search results (something for which we should be glad considering the daily expansion of material available online) and the answer to the question can now – usually – be found within the first few links.

While searching for keywords to answer questions may still be the technique of choice for people who have been using the Internet for some time, a trend I have noticed over the past couple of years, and was reminded of this week while watching a group of Year 7 pupils research a topic, is for users simply to type the question they want answering into the search box and hit return.

Several years ago this would have been a fruitless approach. However, as people expect computers to become more human in their behaviour, the natural language abilities of search engines to understand what users intend by their question has been improving. Indeed, this is linked to the ideas behind the Turing Test in which the responses from a human and a computer are hoped to be indistinguishable, and the ongoing quest to build increasingly humanoid robots. Computers are still not wholly adept at parsing natural language queries, but it will not be long before it is the way we will be obliged to interact with the leading search engines.

As part of the understandable desire to make technology conform more readily to its human masters’ wants and needs, there is the clear link to the development of artificial intelligences with its associated, inevitable, and arguably frightening idea of singularity. In spite of this, I feel there is a still more worrying trend: the aspiration to human-ify technology is in direct correlation to the aspiration of humans to become more machine-like.

Our language choices reflect this and Brian Christian notes in The Most Human Human that fifty years ago, a whizzy new item of technology would be described as being ‘like a computer’, but now we are likely to find ourselves describing a human maths prodigy as ‘like a computer’. For the past couple of years there have been smartphone applications, or ‘apps’, which use readily available content combined with the phone’s GPS-determined location to show sites of interest or businesses. This is a useful tool as the user does not need to enter a long string of keywords or to wade through pages of search results to find, say, a local restaurant. However, recently the apps have developed and also switch the phone’s camera on and superimpose the labels onto the view shown on the screen. The user is therefore looking through the screen at the view they could see if they lifted their eyes, but it is more like a head-up display containing additional information which changes as the camera’s position changes. This is the type of technology that has been seen in fighter jets and the point of view shots of protagonist of the futuristic Terminator films, but is now available on an individual’s phone.

The labelled view of the world is a technologically mediated view of reality presented through – as I have observed here previously – a screen and is known as ‘AR’ or ‘augmented reality’. Users are choosing a technologically enhanced view of their surroundings and thereby negating the need to receive input – a word from the world of electronics appropriated by humans – from other humans, relying instead on their own augmented reality. Humans can have their own head-up display and use the information it presents to follow other humans before them: people who choose to do this are becoming more like the robots on production lines carrying out the repetitive tasks which they took from humans decades earlier.

To me, singularity is therefore not just the evolution – a biological word tellingly appropriated for the inorganic – of technology to reach a point where it is beyond human intelligence, rather it more worryingly reflects humankind’s wish to emulate the futuristic machines of science fiction.


  1. Very quick comment: Searching questions via a search engine also throws up sites like yahoo questions, where hopefully someone has already asked the question in a forum. It's not just because we assume that the search engines are better.

  2. It is very true that results often include hits from Yahoo! Answers but I find the quality of the user generated answers there so varied (even to simple question which could be correctly answered by reading a fully referenced Wikipedia article) I would rather avoid them and spend time reading a more reliable source.

    I don't think it's an assumption of younger students that the search engines are 'better' per se, it is the integration of technology into their lives to such an extent that they feel (doubtlessly unconsciously) there is no need to help the computer by identifying keywords as the technology should be able to parse the natural language for them.

    Also, as part of today's society which demands immediate responses, people seem not to want (or to be able?) to cope with a long page of information and are content with a short, unverified answer. Interestingly, Nicholas Carr comments in 'The Shallows' that the way we read is changing as we look for hyperlinks and flit from one to the next as they are the important bits of the page.

    I deduce that I'm responding to KH here, so I know you are capable of (and happy to) read a long text and sustain your line of thought, but I would suggest that, if you have the opportunity, you look at how other people use websites and read texts.

  3. Whilst I can awe the advantage of AR as a tourist the is also the danger that we are only seeing the AR that the author wants us to see. Going and looking with the human eye is going to be subversive!

  4. Capable of reading long texts, yes. Happy too, not so much. I'm extremely guilty of tab-hopping whilst on the internet (though that's more related to one of your other blog posts than this one), but even when looking something up, I have very little patience online; even to the point of using the "find box" (shortcut: control and f) when faced with a Wikipedia article. But I agree that people feel technology should be able to search using natural language, and that people are content with a short, unverified answer. I think I would add that even when people use Wikipedia which has some verification, very few check it. I know I don't. As soon as there's a number in superscript, I take it as gospel.

    And yes, your powers of deduction are excellent. :) Conan Doyle would be impressed.