Therefore no AI!

WHY A HUMAN COPYWRITER? AI CAN DO THAT TOO! AND IT'S CHEAPER!!!

Many people are currently saying that 'AI' models of the LLM type can significantly simplify and reduce the cost of text production. In fact, many PR and marketing departments - including agencies and even freelance copywriters - are already using this still young technology in their day-to-day operations.

In the case of a text project of low complexity and low standards - in short: of low content quality - this may well work. Webshop operators, for example, can use LLM to automatically create product descriptions. A Wikipedia entry can also be realised in a time-saving manner with an LLM.

However, the situation is different in the case of texts that focus on standards and quality and are intended to convince readers of the expert status of a company and its employees; in other words, that someone really has a 'handle' on a topic, has 'mastered' it - and is not dominated by the topic. Who wants to invest thousands, tens of thousands or even hundreds of thousands of euros in technological offerings from a company that fails to convincingly implement a statement worth reading, an informative study or an article intended to emphasise its own thought leadership? 

On closer inspection, texts created by an AI still contain numerous content errors. Laypeople may read over these and think nothing more of them, but experts - buyers of B2B offers - do. As LLMs are generally fed with data from the entire Internet, the quality of their data input is limited, misinformation is included and misleading formulations are adopted without question. The writing style of an AI also often comes across as wooden, clumsy, more like a thrown-together encyclopaedia entry - a Wikipedia article - than a lively factual or technical article with a relevant - and topical - hook for the respective target group. And finally, LLMs still stray far too often from their given thematic focus. A common thread that leads coherently from A to B to C is usually not found in their work.

Artificial intelligence 

cannot (yet)

• convincingly develop and implement its own text ideas,

• create texts that convince even experts and decision-makers,

• concentrate independently on what is essential for a customer and spin a continuous red thread,

• bring divergent customer wishes down to a common denominator,

• assess the truth and significance of individual pieces of information from the web and sort them out accordingly,

• proactively make suggestions for changes, keeping the topic, magazine, readership, customers and customer parties equally in view.

==> Although AIs can already create texts, they have not yet led their readers to the thoughts and conclusions desired by the customer.