A friend has recently asked me if I think AI writing tools are going to take over our jobs. Since I have played a bit with some AI writing tools available on the market now, my answer is no, I don’t think they can replace human writing. But they can help.
Then, coincidently (or not), I came across an extensive story on AI writing tools published by James Vincent in The Verge – and this article is one of the best combinations of information and reflection that I’ve read in a long time. Now, I really want to share some of its concepts, along with a few thoughts and conclusions of my own.
So, here are five ways how AI can change the world of writing:
1. Writers can use AI to spark creativity
AI writing tools are good for exploring concepts, brainstorming ideas, or producing rough drafts. Like the article’s author, I too have been recently playing with some tools like OpenAI or Writer and I was pleasantly surprised by the naturalness, clarity, and sometimes even creativity of the output.
I’ve used AI writing tools to produce meta descriptions, social media posts, and FAQ sections like the one in this history of Halloween blog post. On other occasions, the output made no sense at all, and all I used it for was to have a good laugh. Which was what I very much needed on some days.
2. AI writing tools still need human supervision
However, they shouldn’t be used to generate text and publish it without having a human revise them. AI-written texts are not reliable, they can contain fake facts, data, or quotes, and the article in The Verge includes an intriguing example. It’s safe to say that AI tools cannot replace humans. Yet.
I don’t see them taking over writers’ jobs, but I can see them being used to improve writing or get you out of a dreaded writer’s block. When I use them, I always read, edit, and re-write the result to make sure it sounds human. And I always make sure to add original elements like stories, insights, or downloadable materials that make each article unique.
3. AI tools can help writers and readers save effort
These tools can also be used for “low-attention” text, as the article’s author James Vincent calls it, meaning something that takes little effort to create and to read. These may be functional texts like emails, documentation, or factual information like trivia.
You may have noticed that Gmail, Google Docs and Microsoft Word are already providing word suggestions as you type. They are built on general word associations and can help you think less when you write – if that’s your goal.
4. Scaling is their best use case right now
AI writing tools are good for scaling simple, boring tasks like summarizing, rephrasing, expanding. You can have a simple conversation with them (think chatbots) as if you would with a toddler. Or you can use them to enforce brand voice in your content design, as I’ve learned from a case study presented by Lisa Jennings Young, head of Content Design at Twitter.
Still, you need humans to be part of the implementation framework from start to finish. From the information that is feeding into the learning model, to approving which suggestions go into the final product, human intelligence is still the key element.
As robot expert Rodney Brooks explains, “every successful deployment [of AI] has either one of two expedients: It has a person somewhere in the loop, or the cost of failure, should the system blunder, is very low.”
5. Beware of misuse
One already noticeable downside could be that large language models may accentuate negative trends like spreading fake information, social biases, or groupthink. When everyone uses auto-complete, it won’t be long before we all start sounding the same.
There are ethical issues arising too, like malicious use from companies (who would have thought?) or individuals. Examples are numerous but can include SEO farms, fake academic papers, phishing bots and so on. When that happens, where should we place the blame: on the machine – or the humans who trained it?
What does the future hold?
To sum up, I would say that no human writer should fear having their job taken over by an AI writing tool. Storytelling is inherently human and while machines are able to generate simple texts, they are unable to understand meaning or language intricacies like irony or sarcasm. They are just “digital parrots”, as James Vincent calls them.
Even so, like other content strategists in Tech out there, I’m excited to see what else these tools can do and how they grow. And curious to hear anybody else’s thoughts on this.