It may seem easy. It may seem cheap. And it may make you popular in the finance department. ChatGPT, artificial intelligence, AI—whatever you want to call it. Is it always the cheapest solution when you need to create great content for your company's website?
Of course, it seems like the cheapest solution because it's easy and quick, and you don't have to hire a copywriter.
But just think about it:
Would you buy a product where the company writes in broken Danish, uses phrases you have never heard spoken in real life, and where you get the feeling that no one has actually put any love and passion into the product?
Perhaps if it is cheap enough, you as a consumer will consider it—but if you are involved in sales, marketing, or are a business owner, you should think about the long-term hidden costs that purely AI-generated website content can have for your business.
Loss of your brand's credibility and authenticity
The thing about AI is that it gives you extremely generic, monotonous, and boring texts. This means you risk your brand voice/tone of voice not being consistent, and you risk not really having one at all.
A recent study shows that your company's credibility drops by up to 43% compared to human-written website content if you use AI to write the content on your website.
Think about it. 43%. It can damage potential customers' impression of your business, and it can create discord among your existing customers.
Artificial intelligence draws on existing content on the internet. You will never be able to get the authenticity, creativity, innovation, and nuanced emotional register of a real human being from a robot. In addition, AI-generated content scores low on Google, as the search engine prioritizes companies that use the E-A-T method (Expertise, Authority, and Trustworthiness).
Hidden tech and organizational costs
Now we come to one of my most important points: the long-term costs of using 100% AI.
Education and licenses
Many people can now write a prompt for ChatGPT and get an output. Very few know how powerful a tool ChatGPT is, which is why companies are currently investing huge amounts of money in AI training for their staff. If your staff are unfamiliar with concepts such as AI and ethics and AI and confidentiality, your company could very easily end up in a legal dispute. But more on that below.
In addition to extra training, implementation, and adaptation of existing work routines, you also need to factor in the cost of licenses for the AI tools you want to give your employees or yourself. It costs money. I have tested many AI tools through my work as a business developer, and the best machines are currently the most expensive ones. The free versions are okay for small tasks, but they don't go very far if you want to use AI seriously for business.
Conflicts among staff
You may have a marketing department, a sales department, and a communications department. They have different agendas—we know that! And there can already be plenty of conflict between those departments. If you introduce AI without limits, you can be almost certain that it will cause turmoil among your staff.
As a rule, a company has fairly clear guidelines for what we can say, how we can say it, and what the brand's tone of voice is.
If you don't have an AI policy, and your staff, or you yourself, start prompting away with ChatGPT, you'll quickly end up with 10 different faces, and your brand will lose all credibility.
So without someone to create clear editorial guidelines, your business can very quickly end up with inconsistent, confusing, and contradictory website content, which can damage your brand's reputation. And that's where the expensive manual corrections and adjustments come in.
Because then you have to pay employees, consultants, or freelancers to fix the mistakes.
Costs only become apparent late in the process
There are many start-up costs involved in implementing AI in the workplace. These include collecting quality data (as opposed to the lightning-fast, generic ChatGPT responses), upgrading infrastructure, hiring and/or training existing staff, and continuously updating AI guidelines.
Customized AI projects can cost millions of kroner, and in addition to companies' average monthly AI license and program costs of DKK 3,000-10,000, you can quickly expect 10-20% extra costs for implementation alone.
And these costs only become apparent in the longer term.
Legal and ethical pitfalls
Finally. Now we've reached a topic that is close to my heart. Namely, the ethical challenges of AI. This is where your company can really step into the spinach. And it can be expensive.
Legal and compliance challenges
Yes, it may sound a bit long-haired, but it's bottom serious.
- AI-generated content may contain errors in disclaimers, terms and conditions, and health information.
- AI can misinterpret legislation and regulatory decisions.
- AI has no moral compass and may treat sensitive and private information inappropriately.
Your company is liable for everything, even if it was your HR employee who ran a document through ChatGPT without carefully reading how the text ended up looking.
Is GDPR being complied with? You should always have a human check business-critical documents and sensitive information if it has been processed by AI. And you should be fully aware of what you are allowing the owners of the AI in question to use. Is it your employees' master data?
Unoriginal content and plagiarism
You must not steal other people's work. Yes, I wrote that. If you blindly use the outputs that ChatGPT or other AI tools give you, it can quickly become a risk to your business. If you or your employees are not source critical—either because they are not academics or you have not provided AI training—you can easily end up stealing other people's intellectual work.
And then your company is most likely looking at a violation of the law. Why do we often talk about AI and plagiarism? It's because artificial intelligence draws on huge amounts of data from the internet, which is existing content. Something that someone has already researched, analyzed, and written. It is super easy to end up stealing other people's work without even knowing it, because many AI tools do not cite sources.
Always ask for a source if you receive a claim as output from, for example, ChatGPT, so that you can fact-check it yourself.
We can look to the US for an example of a well-known lawsuit between The New York Times and OpenAI, in which the newspaper accused the AI giant of quoting their articles word for word.
So there is a lot to consider before choosing the "easy" solution and letting AI take over the work of your copywriter or communications person. I hope this article has helped spark some reflection on when to use AI in your work life and when it is smarter to choose a human being.