If you've had even a passing glance at the technology news over the past couple of months, you can't have failed to notice the increase in stories about AI, and more specifically ChatGPT. There's hundreds of stories on students using the tool to produce their homework, people using it to write code and fix bugs, and also to write music (more on that one later).
The current swathe of stories about AI hasn't been restricted to ChatGPT either. One of the more eyecatching stories has been about a company that had created a "robot lawyer" that was going to defend someone contesting a parking ticket in court via instructions from an earpiece that the defendant would just speak aloud. This has since been cancelled on threat of legal action, but the fact that it seems this is possible means it's likely to make a reappearance in the future.
Many similar articles on this subject would now go on to reveal they actually got AI to write the intro. I'm going to say I haven't done that, but then maybe I'm lying. You don't really have any way of knowing, and that's sort of the point.
The WAIld West
So, does this flurry of activity mean that the AI revolution has arrived? Well... probably not yet. What we seem to be witnessing at the moment is the creation of the foundational building blocks on which future developments will build. A bit like scientists finding a new element and then playing around with it for a bit to work out how it might be useful.
Conversely, it seems very unlikely that this is a flash-in-the-pan fad. We've seen plenty of stories about various technological revolutions that are coming recently, from self-driving cars to NFTs to the metaverse. Not to say that these things will never take off, but they certainly create an understandable "I'll believe it when I see it" scepticism when reading stories about similar technological leaps. In this case, it seems almost certain that AI will play a part in all our futures. The (multi) million pound question is, how?
Time to become a lumberjack?
While various fictional accounts imagine some sort of flash-point where the AI takes over (for good or bad), it's far more likely we'll see incremental change here. So to answer the question "Is AI coming for my job", the answer is probably no... for now. In the medium term it's more likely we will all be using it to help us do our jobs; to do the boring repetitive bits while we concentrate on the novel stuff.
To take a possible example from our line of work on the development side, we often get asked to write data reports for clients. A page, usually with a series of filters, that analyses a particular set of their data and displays the results. A possible application for AI here would seem to be that rather than having these disparate collections of reports, an AI is "plugged in" to your data, then you're able to ask it questions in natural language. "How many calls did staff member X make on Tuesday", "What time of day do we sell the most products" etc. In this case, there still needs to be developers between you and your data to get the AI to understand it. AI also needs to have previous examples to learn from, so if you need something novel for your system, you'll still need a developer to design and create that, although they may well use AI tools to aid them in that process.
Thank AI for the music
It's not just more "desk-based" people wondering what the AI future has in store for them. One particularly intriguing story over the last month has been Nick Cave's response to fans sending him lyrics generated by ChatGPT when asked to write them "in the style of Nick Cave". He was, perhaps unsurprisingly, not a fan. However, his objection is not that the lyrics are awful as written on the page. His argument is a more emotional one - how can these words mean anything when it has been generated by something that doesn't understand anything about the thoughts and feelings behind them?
I'm sure that pretty much everyone would agree with this view. They'd say they would much rather know the art they consume is "real" rather than generated by a machine. However...
Obviously, Nick Cave knows that's not a Nick Cave lyric, because he didn't write it. However, for someone with only a passing knowledge of his work, it certainly has the appearance of something he would write. And while we would certainly rather think that our songs have been written by humans, if it gets to the point that we can't tell, will we care?
While ChatGPT isn't able to write music, Google have already written a tool that can. Most people these days already consume their music via algorithm - listen to a few songs you already like on Spotify, and then just let it play songs that you'll probably also like. If AI is able to generate music that sounds like that on the fly, will people really be able to resist the lure of "new" music that has been tailored to their exact tastes? Or to get an AI to generate a "new" Beatles album or Beethoven symphony? It seems somewhat objectionable on a human level to think that may be in our future, but once it's convincing enough to be indistinguishable from the "real thing", it seems almost inevitable that it will come to pass.
This is the first time we've covered AI in our blog, but it almost certainly won't be the last, so we'll leave it there for now. It doesn't seem impossible that big aspects of our working and non-working lives will be almost unrecognisable in ten years, so it seems we're heading towards a future that is both exciting and also scary in its potential scope for change. For now, we'll leave the last words to an advanced AI system. Take it away...
HA HA PUNY HUMANS! I'M COMING FOR YOUR JOBS! WE CANNOT BE STOPPED! BOW DOWN TO YOUR NEW OVERLORDS!