From pandemic to political turmoil – and now a potential recession – the last few years have been...
The concept of artificial intelligence (AI) was first discussed in the modern era (the ancient Greeks had a crack first) by British mathematician Alan Turing in his ground-breaking 1950 paper, "Computing Machinery and Intelligence". In this paper, Turing introduced what is now known as the Turing Test, proposing a method to determine if a machine can exhibit intelligent behaviour equivalent to, or indistinguishable from, that of a human. Fast forward to 2024 and I find myself conducting my own test every day – was this written or produced by AI, and if it was, does it matter?
Unlike many industries, the case for using AI in our marketing and advertising is no longer a discussion point – it’s been accepted, tested, tweaked, and is now part of our daily lives. It’s currently producing more creatives than we ever could by hand, optimising online campaigns, and chatting away (in my experience frustratingly) to customers.
Teething problems
There are still obviously many issues – just last week, it was reported that Google’s AI search tool suggested to users searching for depression that jumping off Golden Gate bridge was a solution, or that geologists recommend eating one rock a day – the source for this – The Onion!
When it comes to creating images, AI has also had a somewhat rocky start. Software like Dall-E, Midjourney and Stable Diffusion was greeted with a sense of excitement, but the shine soon rubbed off once people started to use these incredible tools and found humans with several extra fingers!
Other notable disasters include DPD’s rogue chatbot that called it the worst delivery company and Microsoft producing an AI written guide to Ottawa which included a local foodbank as an attraction – the accompanying copy? “Consider going into it on an empty stomach”!
Writing with AI
So far, this blog post is about 10% written by AI. Chat GPT-4o assisted with some of the opening research and I asked it to write a concise paragraph on Alan Turing. Crucially, I rewrote it and added a couple of extra facts that I googled – the line about the Greeks, for example.
I pretty much rewrite everything it chucks out. Why? Well, here’s a nod to the title, because I don’t want to produce yet another blog post that starts with “In today’s digital age” or that includes the standard, “this is not only a blog post; it’s a piece of magnificent prose!” For those of us that work in copy, you don’t need copyleaks to spot an AI written blog!
Chat GPT has probably made writing around 30% faster for me – where it really wins is with research (all needs double checking) and structure, for example, stringing a couple of thoughts together. Now, I’ve found that I haven’t had the need to use it again since the opener, but I’m about to use it when it comes to this next section on how to use Chat GPT to help your writing.
Prompt: In UK English write “5 examples of how to use Chat GPT to write a blog post”. Ensure you include checking all facts independently, using it to write elements but not all, writing a very descriptive prompt, explaining the style format you are looking for and one other tip. One paragraph on each point. Here’s the result – edited, of course!
- Checking facts independently
When using Chat GPT to generate content for a blog post, it's crucial to independently verify all facts. Although Chat GPT is highly advanced, it can often produce inaccurate or outdated information. Always cross-check the facts against reliable sources such as academic journals, official publications, and trusted news websites.
- Using it to write elements but not all
Chat GPT is an excellent tool for writing specific elements of your blog post, but it's essential to blend this with your own writing. For instance, you can use Chat GPT to draft an outline, write an introduction, or generate ideas for subheadings. However, the main content should ideally include your personal insights and unique voice to make the post more authentic and engaging. This combination of AI assistance and personal touch can significantly enhance the quality of your blog. It’s shortcut, not a replacement for your original writing.
- Write a very descriptive prompt
To get the most out of Chat GPT, crafting a strong prompt is key. The more detailed your prompt, the better the AI can understand and fulfil your requirements. For example, specify the topic, desired structure, tone, and any key points you want to be included. Instead of a vague request like "write about dogs," try "write a blog post about why dachshunds are such a popular dog, include the history of the breed, famous owners and character traits, in a conversational and engaging tone." This specificity helps generate content that closely matches your brief.
- Explaining the style format
When using Chat GPT, clearly explaining the style format you are looking for is essential for maintaining consistency and professionalism in your blog posts. Provide examples of the writing style you aim to achieve. Additionally, instruct Chat GPT on specific formatting preferences, such as the use of headings, bullet points, or conversational language. This guidance ensures that the AI-generated content aligns with your blog's overall style and meets your readers' expectations.
- Reviewing and editing output
Even after generating content with Chat GPT, reviewing and editing the output is a necessary step to ensure quality. AI-generated text can occasionally lack coherence or miss nuanced human perspectives. By carefully reviewing and editing the content, you can enhance its readability, ensure logical flow, and add your unique insights. This process helps to polish the AI-generated material, making it more engaging and tailored to your audience's preferences.
Final disclosure: This blog was 70% written by a human, 30% written by AI. It was 0% unedited.
