A new AI tool called ChatGPT has been making headlines in recent weeks. Depending on what you read, it’s either the best thing since sliced bread or a dystopian nightmare that will see the robots finally take over.

Some are naturally excited about this technology, its potential, and the way it augments human work. Others are concerned about the role it may play in the proliferation of misinformation online. Meanwhile, marketers and content writers are somewhat fearful for their livelihoods if a ‘robot’ can do their jobs. As with many cutting-edge technologies, however, the reality falls somewhere in between.

What is ChatGPT?

ChatGPT is a Large Language Model (LLM) augmented with a conversational interface, developed by tech startup OpenAI. In layman terms, it’s a complex artificial intelligence capable of predicting how words should be placed next to each other – and a rather accurate one at that.

Drop any kind of written request in ChatGPT’s chatbox, and you will see a complex, structured and fleshed out reply in seconds. But with each idea ChatGPT’s Assistant generates, there are growing concerns around reliability, bias, trust and ethics – not to mention creativity and human emotion.

Is ChatGPT really going to make our jobs obsolete? Is it the revolutionary tool so many believe it is? We’ve spent some time with this new technology to understand what it means for the future of marketing.

The promise

It’s hard to ignore ChatGPT’s capabilities – at least on the surface. Its database was scraped from the web itself, with thousands of terabytes informing its replies. It’s less overwhelming than a search engine, it understands writing style and tone, and it’s great for generating ideas.

Give ChatGPT a topic, some extra details and a very specific command, and it can easily generate a number of ideas. It’s great to brainstorm some creative ideas quickly; most of them won’t work for your specific case, but there are always a few that stand out.

It’s also a useful tool to automate some tedious and manual tasks. ChatGPT can summarise articles, generate keywords, even proofread a podcast transcript to adjust punctuation. Many content creators are using it to draft emails, write scripts and compose outline structures that they can then adapt to their specific needs.

ChatGPT can’t remember every single conversation you ever had with it – but it can remember the one you’re having right now. That means it can reference something you or it said earlier on, so that it really does feel like a conversation with an assistant.

Effectively, ChatGPT puts machine intelligence to the service of our personal and work needs. In the tech industry, this is referred to as ‘hybrid intelligence’ – the combination of human and machine intelligence, expanding the human brain rather than replacing it.

On the surface, this sounds great – but anyone spending enough time with ChatGPT will start to spot some potentially dangerous patterns and perils...

The perils

Alongside the widespread fear among copywriters and content experts that ChatGPT is coming for their jobs, the tool’s users have already spotted some other worrying downfalls.

Reliability, bias, and danger

One emerging issue with ChatGPT is its reliability. Users have reported that the bot’s responses aren’t always accurate, and this is no surprise when you consider its knowledge cut-off point of 2021 – meaning it doesn’t know about events or developments after that time.

The bot is also configured to maximise the similarity between its outputs and the datasets it was trained on. But without an internal mechanism to fact-check its outputs, ChatGPT’s answers are prone to being misleading or, in some cases, totally wrong.

And if you need further proof, turn to ChatGPT’s prompt page itself, which warns that the bot “may occasionally generate incorrect information,” and “may occasionally produce harmful instructions or biased content”. Go figure…

Inaccuracy is one issue, but more concerning are the tool’s issues with bias.

All AI systems carry biases, but some of ChatGPT’s can be harmful: users find it easy to get around guardrails and prompt the bot to share discriminatory language. This is a problem enough in itself, but when you consider the growing reliance on AI in important sectors like healthcare, finance and politics, biased algorithms have the potential to spread inequalities and misinformation at an unprecedented and dangerous pace. There's also a security issue, with cybercriminals beginning to use ChatGP to quickly develop malicious code. Interestingly, it seems to be less skilled hackers who are manipulating the tool – showing the worrying potential for cybercrime to spiral out of control among ChatGPT’s users.

The future of copywriting

And now for the burning question: will ChatGPT replace copywriters?

Not exactly. While the tool has great potential to provide relevant, interesting and useful blocks of text, it’s not yet mastered the art of human emotion – making it pretty impossible for ChatGPT to replicate the headlines, articles and social posts that win awards, stop thumbs, or jerk tears or laughter.

It may get you a pass on a school essay, be competent in writing a generic email, and even pen some basic marketing lines, but ChatGPT lacks the creativity and individuality to produce anything more show-stopping. ChatGPT may give a bad writer the potential to masquerade as a good one in certain scenarios, but it won’t replace the craftsmanship that goes into unique and original content – so copy and content specialists can rest easy and treat the tool like an assistant rather than a replacement.

What's next for ChatGPT (in the words of the bot itself*)

ChatGPT is an ambitious AI tool with a wide range of potential applications, from brainstorming creative ideas to automating manual tasks. However, there are also significant concerns around its reliability, bias and the potential for it to spread misinformation.

Much like previous tools (such as AI art generator DALL•E 2), OpenAI will release new iterations of ChatGPT, and a new version based on neural network GPT-4 is expected soon – which should at least bring a bit more safety and reliability to the tool. But the popularity of AI generators in 2022 has given rise to even more efforts from the wider tech industry, and there’s reason to think that the trend isn’t going away any time soon.

As with many new technologies, it's important to approach ChatGPT with a healthy dose of caution and to always fact-check its outputs. While it can certainly augment human work, we should always be aware of its limitations and potential dangers.

ChatGPT is a promising tool with great potential, but it’s important to use it responsibly. Marketers who approach the tool with this mindset are more likely to use it to augment their work and develop a new specialism – without the fear of ever being replaced by a machine.

*This conclusion was partially created using ChatGPT.

Want to stay in the loop on the latest in social media? Sign up to receive our daily WhatsApp updates, and subscribe to the Battenhall Monthly newsletter here.

‍