Andriana BoyrikovaA golden AI content-generating goose spits out 1000 words in 30 seconds, enabling you to produce ten times more content than you would usually be able to. In a mere two hours, you master typing in the right prompt, slightly editing the output, and hitting publish. ChatGPT, Jasper AI, Writesonic, or any of the hundreds of other similar tools are remolding content creation. At the same time, they are developing at a speed much higher than our ability to understand them. So how can you leverage AI content generators in a smart way? Before you embrace their controversial power, let’s take an objective look at how they can work both to the advantage and detriment of your business.
Although it seems AI writing assistants are the magic bullet for businesses of all sizes and industries, you should remain vigilant. Apart from being extremely complex, AI is developing rapidly and its workings – and consequences – remain inherently obscure. This is why, as a small business owner in a digital world where AI-generated content might become the new normal, you need to ask yourself:
How can I make the most of AI while staying in control?
In the end, you should be in control of your content, not the AI tool.
Who is using AI content generators?
While ChatGPT has captured the public imagination since November 2022 and companies across the globe have been implementing it in their everyday operations, using AI for content generation isn’t a new thing. The Associated Press began using AI in 2014. “With automation, we now follow and produce quarterly earnings reports for 4,000 companies. Previously we covered 400,” Justin Myers from the Associated Press said in a Guardian article. Since 2017, the Washington Post has been using Heliograf, an AI reporter, to cover sporting events that they can’t send a human reporter to.
A Forbes Advisor survey of 600 business owners reveals that 42% leverage AI for website copy and 46% – for personalized advertising. What’s more, according to the AI Marketing Benchmark Report 2023, which summarizes the opinions of more than 2700 marketing agencies and brands, 44,4% use AI for content production.
I have been trying out different AI-powered content generators to determine whether and how I can use them to produce more content that is relevant and valuable to you, my readers. ChatGPT, ChatSonic, Jasper AI, Copy AI, Writesonic, Copysmith AI, Headlime, HyperWrite AI, Rytr, Shortly AI, and Article Forge offer countless opportunities and ease some parts of the content creation process. Despite their differences in features, pricing, and content quality, however, they share the same upsides and downsides.
How does generative AI work?
Generative AI systems, such as ChatGPT, are trained to recognize and recreate patterns of words, phrases, sentences, and paragraphs. They generate text based on what the next most probable word is.
Mike Zhang, a PhD student in natural language processing at the IT University of Copenhagen, explains:
“Generative language models create text based on the data they have been trained on. Essentially, the model can’t go beyond what it has learned. If you want to generate an ad about something that doesn’t exist, for example, a steerless car, and you let the model generate a hundred of snippets about cars, it most probably won’t come up with a description that the car doesn’t have a steering wheel because it hasn’t seen this in the training data yet.
However, they can sometimes appear to ‘go beyond’ their trained capabilities due to the vast amount of text the model has been trained on. They generate plausible-sounding answers, which may not be factually correct. Therefore, critically verifying the sources is necessary.”
So on what data have these models been trained?
ChatGPT was trained on “vast amounts of data from the internet written by humans, including conversations, so the responses it provides may sound human-like.” The team of Jasper AI shares in a blog post that engineers have fed “a wide range of written content from articles and newspaper stories to Reddit threads — which the AI then ‘studies’ and learns to recreate stylistically.” Writesonic “can create new content by learning from the massive datasets provided as examples.” Does it get any more transparent? Unfortunately, no.
The exact papers, books, and other published and user-generated content on which ChatGPT, Jasper AI, or similar tools have been trained remain unknown.
Does this matter? Research1 points out that we trust airplanes and cell phones even though we are almost completely ignorant about their complex inner processes. As artificial neural networks remain fundamentally intransparent, we might as well focus on their safety and performance rather than their workings.
Is it safe to share company information?
Like any other technology, AI comes with a lot of questions and potential risks regarding safety and security. Let’s say you type a prompt in ChatGPT about a product roadmap or sensitive information such as internal meeting notes. Is there a chance that this information ends up on your competitors’ fingertips? Moreover, what are the privacy implications if you share other people’s information without their consent?
In March 2023, ChatGPT users were able to see brief descriptions of other users’ chats from the history sidebar which led to the company temporarily shutting down the chatbot. One month later, Samsung workers unwillingly leaked top secret data while using ChatGPT to help them with tasks.
According to an OpenAI spokesperson on Reddit, the underlying bug was in an open source library.
On the FAQ page of ChatGPT, you will read: “Please don’t share any sensitive information in your conversations.” ChatGPT retains the right to view inputted data to train the system. If you input proprietary information into ChatGPT, it can eventually be ingested into the language model. What about the plugins for ChatGPT that allow it to access recent information on the web? The red team of ChatGPT has found they can “send fraudulent or spam emails, bypass safety restrictions, or misuse information sent to the plugin.” Isn’t this a nudge to opening Pandora’s box? Only time will tell.
Don’t share company information that you don’t want to end up in the knowledge database of a generative AI tool.
Also, before you use any AI content generator, read its terms and conditions and how it uses the information you share.
The unethical sides of AI content generators
1. High stakes of inaccuracy
71% of consumers won’t buy from a company if they lose trust in it. In today’s digital world where any company can say anything, people are becoming increasingly skeptical about what you are trying to sell them including the content you are shoving down their throats. Is your product or service legitimate and is it going to solve their problems? What about the content you publish: is it factually correct?
There is a rising number of instances of ChatGPT combining facts and falsehoods, quoting reputable sources that don’t exist, and generating plausible-sounding answers that are factually incorrect. This is why it is crucial that no matter how believable a piece of content may sound, you need to remain critical and fact-check everything.
2. The plagiarism perils are real
Is it right or wrong to train a generative AI model using authors’, bloggers’, journalists’, or researchers’ work without their consent? Cem Dilmegani, an AI analyst, discusses this in his article: “You can utilize someone else’s data in order to train your generative AI model without any issues. However, what you do with the generated output of this model might infringe copyright law.”
The content most AI content generators create doesn’t come with source credits or citations. Publishing AI-generated content without quoting the experts whose work it is based on and claiming it is yours falls into plagiarism. There is no law against this–at least not for now–so the question is: is this aligned with your company’s values?
Not to mention that if search engines track (almost) the same content in multiple sources including your website, they will penalize you and this may lower your ranking in search engines.
Always run AI-generated content through a plagiarism checker. This way, you avoid unwittingly infringing on someone’s copyright.
3. Disclose or deceive?
Imagine you come across a blog post from a company you are a customer of. It claims to be written by one of their in-house experts. Later, you find out that an AI tool has generated the content. How would you feel about the company?
Deloitte found that trusted companies outperform their peers by 400%. The winning formula to build trust is honesty plus transparency. Being honest with your target audience will gain their long-term loyalty. So if you think you need AI to write an article and you don’t want to lose your credibility, let your audience know that the content is AI-generated. You are being transparent and your readers can decide whether to interact with this piece of content or not.
If companies publish AI-generated content on a massive scale and call it theirs, we will enter what the researcher Aviv Ovadya called the infocalypse: a digital world with readers’ trust reduced to zero where no one knows who and why created the content they are reading.
Four ways AI-generated content works against your business
1. The inhuman touch
Content writing isn’t data. It isn’t pushing words out, either. Content is a form of communication from people for people. Yet, AI generated content doesn’t feel human. On the contrary, it feels like it follows an exact template with no unexpected angle. On top of that, AI can neither tell a story it overheard at a networking event nor share a quote from yesterday’s business magazine article.
Your content is your means of communication with your audience.
Your audience isn’t online profiles and email addresses; they are human beings with real needs, desires, problems, and information interests. For this, you need to know who they are and understand them. And only a human can understand another human.
How can AI evoke emotions which it has never felt?
2. The same content, everywhere
You can use an AI content generator to produce a 2000-word article for your business blog in less than a minute, but so can anyone else. In a digital world oversaturated with bot-created ads, falsified truths, and AI-generated blog posts, you need to make your company stand out.
First and foremost, think about your target audience.
Do they want to read something similar to the cookie-cutter pieces that overflow their social media feeds?
Or are they craving informative content that makes them feel you understand their problems, needs, or goals?
3. Backward-looking ideas
Let’s say, you have hired a professional content writer to write articles for your blog. In her leisure time, she attends a networking event and learns about the launch of a new product, similar to yours. It dawns on her that you can reposition your business so that it attracts a particular niche in the market. Had you used an AI content generator, you wouldn’t have known about this new product let alone considered repositioning your business, which can consequently bring you more customers.
AI content generators can’t use contextual and background information.
Let me give you a real-life example. During a brainstorm session about a LinkedIn ad campaign with a business owner and several team members, our conversation took an unexpected direction and by the end we had come up with ideas for a series of blog posts and case studies in addition to the ad campaign. AI can’t do this unless you prompt it to do so.
4. AI doesn’t care
When an AI tool generates an article for your business blog, it doesn’t care as much about how satisfied you are with it as a content writer does.
Freelance content writers’ bread and butter depends on repeat business. Maybe you have never even considered this but it is central for the quality of the content you receive.
Before you hire a content writer, you have specific requirements and expectations. You do an extensive search on job boards, social media, or in your network, interview the selected candidates, look at their portfolios, and check testimonials. So you build a sense of trust that the content writer has what it takes to do the job. By leaving your content in the hands of an AI content generator, you eliminate all these steps that are crucial for the quality you will get.
Four ways AI-generated content works for your business
1. Enrich your outline and research
Using a tool like ChatGPT can help you with an initial outline of your research: what aspects you should consider or what hooks you might be forgetting.
- Specifically, in my preparation for interviews, I prompted it to complement the questions I had come up with. Often, it generated relevant and non-obvious questions that I hadn’t even considered.
It is also a good starting point when you need to write about a new topic as this is usually the most time-consuming part of a writing process.
- For example, recently I interviewed professors who teach subjects such as computer vision & biometrics or nanoelectronics. Before the interviews, I used ChatGPT to get a basic understanding of each topic. However, when I needed a real depth, I scanned scientific papers.
2. Tons of short copy samples
AI content generators can come up with a myriad of ads, headlines, email subject lines, and potential keywords for a blog post or a web page.
- Once, I spent a whole afternoon puzzling my wits to create 50 ad headlines. Fortunately, AI-powered tools weave in hundreds – in seconds. Often, they are far from perfect but I can easily tweak them to match the respective target audience, tone of voice, style, and purpose. Even better, I can use them as inspiration and create unique ones.
3. Beat the writer’s block
Whether you want to admit it or not, we all have been there: short on ideas for social media or blog posts. While the best source for content ideas remains your customers and readers – their questions, reviews, problems, and needs – AI can be a good addition to it.
- For example, you can use an AI content generator to get 100 suggestions for a blog post. Maybe none is what you really need but these may stir your creative juices and often this is all you need: something to fire up your brainstorming. However, don’t use it as a replacement of what your customers and prospects need and what they tell you in conversations, reviews, or emails.
4. Get the gist
While writing, you can divert from the main topic and discuss points that are irrelevant.
- It helps to use an AI tool in the editing process to summarize passages or strip off embellishments. Eventually, however, you need a human to decide how much should be cut off so that the text still makes sense, doesn’t miss out on the most important points, and sounds human.
Additionally, you can use AI to get the key points from a long piece of content, such as a presentation or an interview.
- You can even extract the most appealing pieces and repurpose them as social media posts or emails. That said, be careful with what information you type in. Do you have the people’s consent to share what they have said during the presentation or interview in a generative AI model?
An AI content generator or a content writer: what is better for your business?
While I was reading the paper “Human- versus Artificial Intelligence2,” one sentence got engraved on my mind: “We are probably not as smart as we think. The amount of cognitive information that we can consciously process (our working memory, span or attention) is very limited.”
So, despite their major drawbacks, AI content generators can assist you in the more iterative, tedious, and time-consuming tasks. Nevertheless, AI isn’t a viable replacement of professional content writers because it is inhuman, generates the same content, lacks forward ideas, and doesn’t care.
- If you are just launching your business and you don’t have a budget to hire a professional content writer, you can use AI content generators and spend a lot of time fact-checking and revising.
- If you are at the stage where you want to make yourself known in the field and build credibility, you should hire a content writer who knows how to use AI content generators to your best advantage.
–
PS: This article was written, from initiation to completion, by a human: me 🙂 .
I hope you have found my tips helpful. How are you leveraging AI content generators? Let me know in the comments.
Featured image: Markus Winkler, Unsplash
Are you ready to hire a content writer? Email me and I would love to help you.
Sources
1. Nosek, B. A., Hawkins, C. B., and Frazier, R. S. (2011). Implicit social cognition: from measures to mechanisms. Trends Cogn. Sci. 15 (4), 152–159; Feldman-Barret, L. (2017). How emotions are made: the secret life of the brain. Boston, MA, United States: Houghton Mifflin Harcourt.
2. Korteling, J. E., Van de Boer-Visschedijk, G. C., Blankendaal, R. A. M., Boonekamp, R. C., and Eikelboom, A. R. (2021). Human- versus Artificial Intelligence. Sec. AI for Human Learning and Behavior Change, Volume 4 – 2021.
3. Venture Beat: Deep Dive in how AI content generators work
