AI can do all sorts of things - but credibility needs people

Fundamental - in practically all phases of the work process. We can optimize many things that have always been part of our editorial work; some processes can be automated. We have always been data-driven, but now we can make workflows even better and more profitable. The challenge is that artificial intelligence gives the impression of being able to solve everything - the real quality lies in turning quickly generated output into special products.

First and foremost in planning, research and finding topics. When we tackle a new topic, AI is our sparring partner. It helps us to sift through exciting sources, cluster topics and make initial deductions. The combination of AI and data interfaces is also interesting: AI checks previous content performance, performs media and potential analyses, conducts social listening, identifies recurring topics in forums and comment columns and scouts trends and suitable experts.

We are developing our own AI agents. These are specialized assistants that have a specific job: One has ideas checked by synthetic target groups, others format texts or take care of visibility optimization. The key point is that these agents don't just do the work for us, but we train the models so that they become real dialog partners, for example in idea generation. It's a bit like training a new colleague. It takes time at first - but once you have invested this time, they can really take the pressure off us and improve our results.

This is probably the key issue for editorial teams today. AI is revolutionizing journalism and content marketing, but without a clear stance, we risk losing our credibility. In some projects we use AI extensively, in others we limit ourselves to research and accompanying data analysis. This is always done individually and in close consultation with our customers and their specifications, including with regard to internal company regulations. Human control and quality assurance remain non-negotiable. It is therefore important to provide appropriate advice, take fears seriously, look for suitable approaches and use AI responsibly.

Absolutely not, because hallucinations are a real problem - AI sometimes invents facts that sound plausible but are simply made up. And it often misinterprets contexts. Then there is the bias problem: AI systems often unconsciously reproduce prejudices from their training data. This can be subtle - a certain choice of words that excludes population groups, or one-sided perspectives. AI-supported systems can produce impressively formulated texts, but sometimes they also produce completely useless texts. Content errors or distortions are often hidden in passages that are convincing at first glance. This is why consistent fact-checking is so important. If you leave the responsibility for the text solely to the AI, you will quickly get into trouble.

In addition to topic development and production, technology, context and the associated effective content strategies are becoming increasingly important. This is why comprehensive consulting expertise is necessary, including on the processes. There is no question that AI expertise is now an absolute basic competence. But this makes human quality control and creativity more valuable than ever. The more automated content is created out there, the more what only humans can do counts: telling real stories, creating emotional connections, penetrating complex contexts and making ethical decisions.