AI and Content Writing
There are reasons why it’s illegal to marry your cousin and one of those is because of the lack of diversity in the gene pool leading to abnormalities in any offspring. An odd way to start a blog on AI but it’ll all make sense!
Whilst AI may be a cost-efficient time saver for business, it feels like a threat to creatives as it can seemingly do the work that we do in a fraction of the time and at a fraction of the cost. Clients have come to me and said, “Chat GPT has created this for me, can you do better?” Well, my answer is yes and here’s why.
• Avoids ‘inbreeding’
• Originality
• Relevance
• Accuracy
Generative AI has a small ‘gene pool’
We’ll start with the inbreeding issue, not a sentence I thought I’d ever write in a blog. When you ask AI to create a piece of work for you, it’s referencing what’s already on the internet. ‘That’s ideal and exactly what I want,’ I hear you cry, ‘After all, I was going to perform all my research on the internet so this will save me time and money!’
But there’s a problem, and this is where inbreeding becomes relevant: if more and more of the internet is made up of AI generated content, then what’s produced for you has fewer unique sources, is less diverse and as a result, it will have abnormalities.
My SEO guru friend Dominic Potts of Full Pivot likened it to a photo which is photocopied and then the photocopy is photocopied and so on which is a great analogy. If you can remember using photocopiers, photocopying your classmate’s handout always resulted in a handout of slightly less quality than the original and considerably poorer quality than the source itself. There are some other fantastic examples in this New York Times article (‘When A.I.’s Output Is a Threat to A.I. Itself’), which goes into more detail about this concept too such as considering the above in the context of a medical paper: if the ‘gene pool’ is small, the piece of work will be less accurate and incomplete, and potentially lead to an incorrect diagnosis or treatment.
OK, let’s bring it back to what this means for you.
AI is not wholly original writing
Search engines are becoming wise to AI and the age-old issue of badly written content stuffed with keywords, like blogs and articles, and are tweaking their algorithm and search results accordingly. Google released a change a few months ago which is deliberately looking for quality content that has been created by humans for humans to read and put to practical use.
Image by Andrea de Santis via Unsplash
Humans writing for humans mean that your business writing has a personal and, importantly, a unique touch. AI might know about your business and award-winning service but not about the client who loved what you do so much they sent you a case of champagne (that’s me manifesting!); bots will be able to trawl the internet to find information on your best-selling, revolutionary toe-less socks (obviously made up!) but won’t understand the story behind their creation – the kind of story which consumers love. Sure, an AI blog will mention socks a lot but it won’t have access to that unique, engaging story that converts into sales.
Relevance
It’s also worth noting that some AI models can only access data up to a certain date. For example, for GPT-3.5, ChatGPT's knowledge base has a cut-off date of September 2021or January 2022 for GPT-4. What this means for you and me is that ChatGPT can't accurately handle requests related to events that occurred after that time.
This is really limiting! Most of us want to read about issues and topics that are relevant to our lives today, comprised of accurate up-to-date information. A lot can change in a couple of years!
AI cannot check its own work
A writer will pay attention to detail and strive for accuracy and this may not be the case with AI. My recent particular interest in AI was piqued by an article I came across from investment bank, Goldman Sachs featuring their head of stocks research Jim Covello. In a New York Times interview it was stated that Covello found that, “generative artificial intelligence, which can summarize text and write software code, made so many mistakes that it was questionable whether it would ever reliably solve complex problems.” You can read the New York Times piece here (‘Will A.I. Be a Bust? A Wall Street Skeptic Rings the Alarm') or the (free, no paywall) Goldman Sachs source paper here.
Image by Emiliano Vittoriosi via Unsplash
AI is good for …
AI does have its uses of course and Dom was telling me how many people are now using it as a research tool. I have to admit that I have used it like this when looking for a particular product or wanting to learn more about a subject. The search engines make it almost impossible not to use it, don’t you find? The AI list comes up alongside your search results in my experience, unless you use a search engine like Duck Duck Go.
AI apps are available and used by organisations to write up minutes and provide summaries which saves time. But they’re not yet 100% trustworthy and still require a check over to ensure they’re correct.
This may seem like a slamming of generative AI but its advance is all but inevitable in my opinion (despite what GS’s Covello says, “building too much of what the world doesn’t need ‘typically ends badly.’”). I think everyone knows it has its faults but the lack of diverse original sources – the inbreeding – was one I hadn’t considered, and I thought it was important to highlight. Thanks Dom!
Remember that when you employ a freelancer like me rather than a faceless tech giant, you’re also supporting the small business economy which ripples outwards into local communities and onwards.
And, of course, no AI was used in the writing of this blog post.
If you would like to employ me to write original, accurate, relevant and interesting pieces for you, please contact me Contact — Caroline Kings