It’s no secret that I’ve been researching and learning about making AI for a number of years. Through that time a whole bunch of questions occur to me and I scribble them down to pursue in that magical time that rarely arrives known as ‘later’.
To purge some of these queries, I am gathering up a series of short bits about AI that get me thinking. There might not be brilliant answers of course, but I think they’re worthwhile as themes as we carry on through a bit of a peak in artificial intelligence research and writing.
Not the Terminator – then what?
Working in tech for a long time now, I come across this issue regularly in reading and research. Here’s my opening reaction – I dislike seeing the Terminator being used to represent AI. Also things that are a bit pesky to me include glossy white robots, bionic arms and hands, people in hoodies coding in a dark room and the favourite – the Matrix graphics raining a waterfall of numbers on our lives like a kind of semi-transparent wallpaper.
Don’t get me wrong, I’m a huge sci-fi nerd, I love the thought exercises, the magic, the deep research that goes into a lot of these works and the good fiction really gives me plenty to think about in terms of real-world scenarios. A little perspective on this – The Matrix came out in 1999. The Terminator was released in 1984. So why do we get stuck with the same visual vernacular?
I work in media (Let’s make it clear that my academic pursuits and writings in no-way represent who I work for or who I have worked for in the past.) and I get it. I do.
Part of the issue is finding an audience. To attract people to read about your amazing AI revelation and important update on the development of AI that you might feel should affect lots of people – you could probably do with a real hum-dinger of an image to sell that story.
I’m not talking about academic papers – which are not usually illustrated apart from graphs.
Artificial intelligence is at a stage where it’s available in different sectors and forms – so it’s possible not to represent it as a robot, or as lines of code. So I come down on the side of finding something that closely matches the section of the topic that is being described. But when that’s some deep down maths, is it still appealing for public consumption? Is it recognisable as AI?
There are of course going to be cases where the examples listed above are perfect for certain articles. And probably a relief to the author who needs to show things clearly and in a way that sparks the interest of readers.
Earlier this year in a guest book at a place in Rovaniemi, Finland, I came across something that blew my mind – mostly because I have drawn some myself. Maybe for a lack of paper, someone had drawn the outline of a neural network. In that way that your own perceptions create your world, I found this amazingly personal and magical. I would love to know who wrote that and left it there.
I’m equally aware that showing those pictures to people who work in other fields would look at it and shrug. Meh, some kinda maths right? If I walked in and found a cute little robot that had some neat natural language processing tricks and learned from me each day how to become a welcoming hostel host, that I suspect would be interesting to a lot more people.
It’s basically ‘sexier’ at the moment to show a robot to illustrate certain kinds of computer science. I am guilty too of probably being quicker to click when I see an interesting robot I know about. Also that if someone shows me say….physics equations I don’t understand, I might not engage as easily.
Maybe it will take time or maybe the deeper technologies just don’t need to be illustrated. You don’t need an image of a computer program to pique your interest in driverless cars, probably just a picture of the car it is in.
I’ll keep thinking on it I guess. But for now that’s what I’m thinking on AI, along with a few other things that I’ll probably post over the coming weeks.
Thanks for reading.