AI and the Future of Work

The explosive unveiling of ChatGPT at the end of last year set the professional world alight with questions regarding the role AI will play in the future of work. For many in the creative industry, the instant reaction was one of fear and despondence. Everyone knew that artificial intelligence would one day reach a point where it was able to replicate human language to such a standard, but no one saw it coming this quickly. As such, many creatives naturally entered survival mode, perceiving the proficiency of this technology as a declaration of war on their livelihoods. This reaction is both rational and irrational.

ChatGPT (or one of its many language learning rival platforms) will change how businesses work, especially with regards to customer-service and creative output. And there will be job displacement, as well as a great deal of retraining. That’s the downside. The upside is that we’ve been here before with other industries (and those whose roles are now in jeopardy were conspicuously less vocal when it was blue-collar jobs on the chopping block), and we found a way through. Spellcheck did not mark the end of proof-readers, even the most automated factories still require human supervision, and, though there’s not a human alive that could beat a robot at chess, the game lives on and is enjoyed by many around the world. To re-work an iconic line from a film franchise – one that explores the potential dangers of human folly in innovation – life finds a way.

What is ChatGPT?

Perhaps we’ve gotten ahead of ourselves. For those who’ve been living under a rock these past months, it’s probably best to start by explaining what this platform everyone’s been making such a fuss about is and what it does. And who better to explain that than the platform itself? When asked to describe its purpose and function, ChatGPT states:

ChatGPT is a large language model developed by OpenAI. It uses deep learning techniques, specifically the transformer architecture, to generate human-like text based on the input it receives.

The way ChatGPT works is by taking in a prompt of text as input and then generating a response based on the patterns it has learned from the vast amount of text data it was trained on. The model has been trained on a diverse range of texts from the internet, including web pages, news articles, and books, among others. As a result, it has acquired a broad understanding of language and can generate a response in various styles and on many topics.

Essentially, it gives human-sounding responses based on a prompt, pulling its answers from the recesses of an expansive database. The more generic the prompt, the more generic the response and vice versa. Upon its release, The Times declared it “the world’s first truly useful chatbot.”1 Meanwhile, VentureBeat said generative AI was “revolutionising how we experience the internet and the world around us.”2 High praise. But is it warranted?

What are the benefits?

The benefits of ChatGPT and its generative AI counterparts are immediately clear. It produces, in a matter of seconds, answers on pretty much any topic, in whatever form the user requests, be that article, social media post, etc. Not only that, but the writing is consistently functional, if still a way short of being anything more than that. For basic written tasks, it’s undeniably a time-saver. In a customer service capacity, for example, it can instantly access and collate a vast pool for information and provide it to a customer in need in real-time, improving customer satisfaction and freeing up the customer service operator’s time for potentially less menial tasks.

Similarly, DALL-E2, ChatGPTs visual equivalent, also made by OpenAI, has stark and immediate benefits. For someone looking to add visual flare to a presentation, they can simply type in a prompt for the sort of imagery they want and have something ready in seconds, as opposed to hiring a designer or illustrator, which is more time-consuming and costly. Or they can show an AI-produced image to their design team before a project begins to give them a better idea of the kind of assets they’re hoping to receive.

What are the disadvantages?

Obviously, there are drawbacks too – for employers and employees alike. OpenAI itself acknowledges that ChatGPT has “the potential to occasionally generate incorrect information or biased content,” as well as having a “limited knowledge of events or knowledge after 2021,”3 the year the model was trained. Galactica, Meta’s own large language model, was taken offline after just three days. In that brief spell, the model was found to be unable to distinguish fact from falsehoods and would fabricate papers (that it would attribute to genuine authors), including generating wiki articles about “the history of bears in space.”4

Generative AI like ChatGPT is also thought to be susceptible to cyber-attacks, whether that’s the spreading of malware or gathering of personal data for phishing scams. That’s not to mention the more dystopian possibilities. One user asked ChatGPT itself what the worst possible outcome of its implementation would be for human employment. Its response? “The worst possible outcome of human employment of ChatGPT would be if it caused widespread unemployment and social upheaval, leading to societal collapse.”5

There’s a bleak hilarity to the model’s casual acknowledgement that it may help bring about the end times, but it points to genuine concerns regarding the job market, where its impact will be felt acutely. Once-stable careers could well fade from view, and quickly. As cashiers and manufacturers discovered some time ago, automation leads to upheaval. This time it will be on a scale previously unimagined. The implications for businesses are obvious – how many cuts? Where? What level of loyalty do they owe their employees? On a more macro level, governments around the world will be forced to contend with the possibility of thousands, potentially millions, of previously stable members of the workforce finding themselves in the unemployment line.

The less doom-laden news is that it won’t happen yet. The software is hugely impressive but its writing to date lacks spark, meaning that while it may suffice for basic, more turgid corporate work, for anything more creative it has a way to go. But the next iteration will improve, and with its increased sophistication, these issues will increasingly come to the fore.

What next?

Realistically, the only option is to embrace AI in some form or other, lest we face the fate of the last silent film stars, unable to find their worth in a world of sound. For as long as there has been technology, there have been technological advancements. This is a seismic one, but not necessarily a death knell or guarantee of future obsolescence. As noted earlier, ChatGPT’s output is only as good as its input. The ability to maximise the quality of that output through a smart use of prompts will soon become a pivotal skill. And for so long as the quality of the software’s writing remains competent but limited, adding more interesting layers to quickly generated AI jumping off points may prove key too. Though there are pitfalls to that…

AI and creativity

The acclaimed author Ted Chiang recently offered his views in the New Yorker. One of his primary concerns was the impact on originality, particularly regarding writing. “If you’re a writer, you will write a lot of unoriginal work before you write something original,” he said. “And the time and effort expended on that unoriginal work isn’t wasted; on the contrary, I would suggest that it is precisely what enables you to eventually create something original.”6

More specifically, on the idea of ChatGPT’s potential role as a tool for kickstarting ideas that could then be sharpened and expanded on, he writes:

Your first draft isn’t an unoriginal idea expressed clearly; it’s an original idea expressed poorly, and it is accompanied by your amorphous dissatisfaction, your awareness of the distance between what it says and what you want it to say. That’s what directs you during rewriting, and that’s one of the things lacking when you start with text generated by an A.I.

Legendary musician Nick Cave goes one further when discussing AI-generated music. “It could perhaps in time create a song that is, on the surface, indistinguishable from an original,” he says, “but it will always be a replication, a kind of burlesque.”7 Before adding, in potentially the most Nick Cavey of all Nick Cave quotes, “Songs arise out of suffering…data doesn’t suffer.”


The truth is that, as of right now, we don’t know quite how impactful AI tools like ChatGPT and DALL-E will prove to be. But the smart money is on very. As such, it’s worth taking time to understand the software now, at least in some capacity, so as not to be blindsided when more sophisticated versions of the software emerge in the future to truly shake things up.

AI will feature increasingly prominently in day-to-day work. This may cause professional displacement, or potentially it will make jobs we’ve long taken for granted more specialised. As we have learned in other industries, there will always be a market for the hand-crafted and bespoke – the homemade mixtape your friend tailored for you will always have more charm than the playlist the algorithm made. The future is uncertain, but as George Bernard Shaw said, “Progress is impossible without change.” Let’s hope the inevitable change AI instigates will take us forward.