Drofa Comms Content team conducted an experiment: we let our colleagues from different departments analyse three incognito texts on the same topic — the role of AI in PR. One text was written by a human, a PR professional to be precise, and two were generated by AI. Here is part I, where we share the results of this experiment.

As we can see, with a given clear requirement specification, AI can work its magic and provide a decent piece of written material. However, a human being is still needed to compile the assignment, check it and edit it to meet the standards. 

Who knows, maybe in the next couple of years, we will see a rise in job descriptions replete with “ability to work with AI”. For now, I would like to summarise the main points that will help writers and other content creators get the most out of working with AI.

How to Work with AI When Writing Texts

Set tasks correctly

AI is great at finding the best-matched answers to your questions from the web, but it cannot generate content from the void. It works with the data and patterns it has already learned, so the quality of the result depends directly on how clearly you formulate your request. If the task is vague, the answer will be vague too. That’s why you have to be as precise as possible.

You need to assign tasks correctly, including all the necessary input data: subject matter, exciting topics, word count, tone of voice, target audience, and format. For example, there is a big difference between asking AI to “write an article about marketing” and asking it to “write an 800-word blog post about social media marketing trends for small businesses in 2023 in a professional but friendly tone.” 

It is also helpful to mention specific points that must be covered, keywords that should be included, or examples you want AI to follow. If the text is meant for a particular region or industry, specify that as well. The more context you provide, the better the AI understands your expectations and matches them

Edit to the max

As we can see from the experiment, even with a given word count, AI can misinterpret your words or make a mistake. Be careful with logical transitions and opinions stated in an AI-generated text. Any piece should be edited and proofread, and AI is no exception to the rule.

Be careful with logical connections and the opinions stated in an AI-generated text. Sometimes the arguments may sound convincing but lack depth or evidence. In other cases, AI may unintentionally contradict itself within the same piece. These issues are not always obvious at first glance, so careful proofreading is necessary.

Also don’t forget to check the structure: does the introduction clearly present the topic? Do the paragraphs flow naturally? Is the conclusion strong and consistent with the main points?

Grammar and spelling also require attention. AI usually produces clean text, but there can still be minor awkward phrases or unnatural wording that appear. In the end, AI can generate a draft, but only a human can polish it to a professional level.

Check for plagiarism

AI basically generates content based on existing texts on the Internet. It does not intentionally copy specific articles, but it relies on patterns found in publicly available texts. Because of this, there is always a potential risk of similarity to already published materials.

And the risk of plagiarism varies depending on the AI model you use and the topic you choose. 

So before publishing, I do recommend running the text through a plagiarism checker. Having high originality results will protect your reputation and save you from unnecessary procedures. Even unintentional copy can negatively affect credibility and ruin the trust you have built for years. 

The more original value you bring, the lower the risk of plagiarism and the higher the quality of your work.

Fact-check the numbers

Not only do humans generate false news, but machines are also pretty self-sufficient in that matter. Let’s avoid spreading disinformation and check the facts before publishing.

Always verify numbers, dates, names, and references using reliable sources. This is especially important for sensitive topics, such as health, finance, politics, science, or law, where inaccurate information can have serious consequences.

Do not assume that a well-written paragraph guarantees factual accuracy. Treat AI-generated data as a draft that needs you as an editor to confirm it. Fact-checking protects not only your readers but also your professional reputation. 

Add a human touch

AI can lack originality, and search engines’ ranking systems aim to prioritise content that demonstrates it. If you want to rank high in search results, be sure these criteria are met. Revise the content by adding unique insights or your own opinion and excluding similarities.

Moreover, a human touch makes the content more engaging and relatable. Readers appreciate the authenticity and emotional part of your text. Even small details, such as a short story, an expert opinion, or a practical tip, can significantly increase the value of the text. People love texts that somehow touch their hearts.

Sometimes, even the greatest story can be ruined with AI’s cold and machine language. Try combining AI efficiency with human creativity, because AI can help save time and structure ideas, but only humans can bring individuality.

Conclusion

AI solutions are rapidly evolving, and we are always looking for the next efficient tools to implement and test in our work. These technologies open new opportunities for faster and more productive content creation, and we shouldn’t lag behind.

However, successful collaboration with AI requires responsibility and attention to detail. Clear task setting, thorough editing, plagiarism checks, fact verification, and a strong human touch are all essential steps in the process.

Stay tuned for the updates, and always go hand in hand with the market.