Jack Lenton, Senior Content Executive at Embrace ponders the impact of ChatGPT and AI on content generation.
Every so often, an invention comes along that is so ground-breaking it completely transforms the way we live our day-to-day lives. In the 20th century it was cars and aeroplanes, and more recently, the internet and smartphones.
As someone who has spent their entire career so far primarily creating content of some description, I have heard more than a few murmurs that AI technology, and more specifically ChatGPT, is about to be the next of these. How true is this?
Could ChatGPT have written this blog?
I have to admit, the first time I used ChatGPT my jaw hit the floor. I watched in awe as, after being given a simple prompt, it generated in seconds a piece of content that it would’ve taken us – the puny humans – far longer to create. Just for fun, I asked ChatGPT to write this blog. Here is how it started things off:
Pretty good, right? It continues for another 300 words or so in near-flawless prose, making a number of valid points and tying it all up neatly with a conclusion. It’s enough to have any content creator quaking in their boots – but should it?
Advantages and limitations
It’s my belief that content creators should see ChatGPT as an extremely useful tool to improve the efficiency of their own writing, rather than a threat to their existence. Unsure how to finish a sentence, or need some quick copy for SEO? ChatGPT is a godsend for this.
However, there are also a number of limitations to be aware of. It currently writes in American English and struggles greatly to chat in other languages. It has only been trained until 2021, and being unable to access the internet, it occasionally gets facts on current affairs very wrong, which could be embarrassing if published unchecked. It also has a penchant for repeating a few select phrases across a wide array of content, which doesn’t make for interesting reading.
Moreover, the creators, OpenAI, have imposed sensible profanity limitations to prevent users from tricking the AI into spouting bigoted nonsense, but this also has the capacity to go wrong – it recently refused one user’s request to write a speech on why carrots were the worst vegetable.
Perhaps a better question to ask then, is just how much potential does this technology have to develop and overcome its limitations?
Is AI our new Deep Blue?
In 1997, the supercomputer Deep Blue beat world chess champion Garry Kasparov in what is now considered a milestone in the history of artificial intelligence. Deep Blue is the size of a refrigerator. Today, put the chess app on your smartphone on its hardest setting and it will comfortably beat any human, even grandmasters.
So are AI programmes like ChatGPT already nearing the peak of their development, or are we only at the ‘Deep Blue’ stage of their capabilities? It’s hard to imagine how exactly it could progress further, although the early signs are there. It has already been taught to write surprisingly complex programmes in code – the popular YouTuber and gadget geek Tom Scott got it to write a code for automatically organising his email inbox. Meanwhile, other AI programmes like DALL-E are creating incredible photorealistic artworks of all descriptions, posing a huge threat to very talented human artists and designers.
Right now, there are probably only a few boffins at Google and Microsoft who can make any kind of fully-informed prediction about just how much this technology will eventually be able to accomplish, but there is room for excitement and apprehension in equal measure.
Is this content creation’s Napster moment?
The popular file-sharing application Napster posed a major threat to the music industry in the early 2000s by allowing users to share their favourite tunes with each other without needing to buy them. Exasperated record label bosses couldn’t abide the idea of people simply listening to music on the internet for free, and after a wave of lawsuits, it was promptly shut down. There were many who thought that would be the end of it.
Today, you would be hard-pressed to find anybody who doesn’t access their media through streaming services such as Spotify. Try as they might, industries are rarely able to stem the tide of technological advancement, and those who don’t move with the times tend to get left behind. For every Blockbuster, there is a Netflix waiting in the wings to take its place.
AI writing programmes like ChatGPT may well be the first nail in the coffin for some industries, but new ones will probably also be created around it. Proficiency at ‘talking’ to an AI is likely already a skill that many companies are starting to look out for in their hiring processes. To put it bluntly, evolve or be extinct!
Should we be scared?
Harlan Ellison’s 1960s post-apocalyptic science fiction short story I Have No Mouth, and I Must Scream tells the story of a super powerful artificial intelligence that brings about the near extinction of humanity.
The AI in that story tires of organising the humans’ constant wars with each other, and eventually decides that the world would be better off if humans didn’t exist at all.
While we might not have to worry about ChatGPT murdering us all just yet, there are legitimate ethical concerns on the horizon. The ‘dead internet theory’ posits that online blogs, comments and chats could one day be made up of almost entirely AI content, leaving the small proportion of human users in a constant state of questioning whether any of what they’re reading came from the mind of a human or a computer.
Programmes like ChatGPT have brought what was previously a fun thought experiment one step closer to a grim, Orwellian reality of post-truth. Although chatting to an AI is incredibly useful and even fun, it’s vital to ensure that we don’t squander our connections with each other in the process.
Welcoming our AI overlords
So how would an AI conclude this blog? Over to you, ChatGPT:
Good point! But then again, it would say that.
This latest thinking article was written by:
Jack Lenton
Senior Content Executive