Between Efficiency and Generic Content: AI in Social Media Marketing

When I read posts on LinkedIn, I notice something more and more often: many of them sound almost too perfect. After the first few sentences, I often find myself thinking: This post was written with generative AI. A moment later, I lose interest and keep scrolling.

I don’t think there is anything wrong with writing with ChatGPT. Quite the opposite: I use AI myself, and I believe it can save us a lot of work. But when everything suddenly starts to sound the same, something feels off. In the worst case, the message does not get through at all because people simply stop reading. This made me think: How can companies use generative AI in social media communication while preserving the authenticity and credibility of their messages?

AI is already widely used in social media communication. According to the Social Media Study Switzerland 2026 by ZHAW and Bernet Relations, 86% of the large Swiss companies and organisations surveyed use AI tools. They mainly use them to write, edit and summarize texts and for translations. From the outside, however, it is almost impossible to tell reliably whether a post was written with generative AI. Current AI detection systems can also easily produce inaccurate results, particularly with edited texts and unfamiliar language models.

More content with AI does not automatically mean more relevance

AI can make content creation significantly more efficient. When writing posts, it can develop variations for different target groups, shorten content, translate it or adapt it for another channel. I see this as a great opportunity. In my previous work in corporate communications, it would have saved me an enormous amount of time in many situations.

But more content does not automatically mean more relevance. And an efficiently produced post is not necessarily good social media communication. If everyone can produce more content more quickly, the first result is simply more noise: more posts, headlines and messages competing for our attention. In my view, AI does not automatically make it easier to be heard. It may even make it more difficult.

Standing out on social media therefore takes more than a well-written text. Good social media communication needs creativity and the courage to put your own voice out there. Ideally, that voice is part of a brand’s communication strategy. This also requires a clear decision about which tasks AI should be used for – and which ones it should not.

“On brand” is not automatically authentic

Strong brands have brand guidelines that define how they communicate, which values they convey and how they present themselves visually. When these guidelines are integrated into an AI system, it can produce texts and images that fit the brand remarkably well. AI can learn which terms a brand uses, how it addresses its target groups and which wording or visual themes suit it. This helps create consistent and recognizable content. But does that already make the communication authentic? Does a text like this automatically feel credible?

I believe that good prompts and clear brand guidelines can primarily help brands create consistency in AI-generated content. But a consistent tone of voice is not the same as authenticity. Communication only becomes authentic when a company lives what it communicates externally. Content also needs to be reviewed by a person who ultimately takes responsibility for it.

Of course, companies could communicate things that did not reflect reality long before AI existed. Generative AI did not create this problem, but it definitely makes it easier to produce large amounts of content in a short time that sounds and looks “on brand”. Ultimately, a social media strategy is part of a wider corporate strategy. Whether communication feels credible also depends on how a company is led and the culture it lives in its day-to-day work. Values and visions cannot just sound good. They also need to be reflected in the company’s actions.

Practical recommendations for companies

But back to the question at the beginning of this blog post: How should a company use AI in social media communication so that its messages reach its target groups while conveying credibility and authenticity?

It is worth taking a brief look at the research, even though the field is still young. Initial findings show that AI can improve the quality of individual results. At the same time, however, the ideas and texts that emerge become more similar to one another. The studies do not focus directly on social media posts, but they remind me of what I currently see on LinkedIn. A company’s key task should therefore be to keep strengthening its brand through social media communication that is distinctive and credible.

Based on this, I see several key recommendations for companies:

• Delegate routine tasks to AI: Transcriptions, subtitles, translations, formatting and adapting content for different channels can largely be automated. Even with routine tasks, however, editorial responsibility before publication remains with the communications professionals.

  
  

• Ensure brand consistency: As soon as generative AI writes or edits texts, it needs clear guidance on the company’s language and tone of voice. Brand guidelines help ensure that a text fits the brand. But that alone does not make it authentic. Authenticity also requires a company to act in line with what it communicates.

  
  

• Use AI as a sparring partner: For research, topic analysis and repurposing existing content, AI can provide an initial draft, introduce new perspectives or suggest different formats. The results and sources need, of course, to be checked.

  
  

• Let AI analyze data, but do not delegate decisions: In social listening, AI can analyse large numbers of posts, organise topics and highlight noticeable changes. This can be particularly valuable in crisis management. But a person still needs to decide what a negative comment means for the company and how it should respond. As a general rule, the more sensitive and consequential a message is, the more important human judgement and responsibility become. AI can also add enormous value to pure data analysis, for example when a company wants to assess how well its social media strategy is meeting its defined objectives. But the data provided to the AI must first be checked for accuracy.

  
  

• Keep a human voice in personal communication: This is particularly important for posts published in the name of a specific person, such as a senior leader. AI can create a draft or shorten a text. But the person responsible should still ask: “Do I really think this? Would I also stand by it in a personal conversation? What would I never write this way myself?” A distinctive voice does not require every sentence to be typed personally. What matters more is that the thoughts and position come from the person and that the text is carefully reviewed before publication. Sometimes the better text emerges from a conversation with a communications professional. And sometimes the easiest option is simply to open the laptop and write it yourself.

  
  

• Train employees: Corporate influencer programmes depend on employees sharing their own perspectives on a company. If they use AI, they also need to know how to preserve their personal voice and handle information responsibly. Otherwise, there is a risk of losing exactly what makes these posts valuable: a personal insight told in personal language.

  
  

Decisions about the use of AI in communication belong in the social media strategy, in expanded brand guidelines and in discussions between communications, legal, data protection and the relevant specialist teams. Companies also need to discuss when the use of AI should be disclosed.

We do not need even more perfect content

Ultimately, every company needs to observe and evaluate its own results: Which posts work well? Which ones reach their target groups, and which ones disappear without being noticed? The use of AI can have very different effects depending on the context.

For me, one thing is clear: The content that excites me most on LinkedIn is content that does not sound like AI. I don’t think we need even more perfect content on social media. We need people and brands that stay true to their own language, even when they use AI.

Sources

Transparency note: I originally wrote this article in German and used ChatGPT to translate it into English. I also used generative AI for research, idea generation, and language editing. The ideas, personal perspective and final edits are my own.

ZHAW & Bernet Relations (2026): Social Media Study Switzerland 2026. Around 100 large Swiss companies and organisations were surveyed. View the study at Bernet Relations

Dugan, L. et al. (2024): RAID: A Shared Benchmark for Robust Evaluation of Machine-Generated Text Detectors.Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics. The detection systems studied proved to be insufficiently robust when faced with modified texts, different generation methods and unfamiliar language models. View the study

Doshi, A. R. & Hauser, O. P. (2024): Generative AI enhances individual creativity but reduces the collective diversity of novel content. Science Advances, 10(28), eadn5290. View the study

Anderson, B. R., Shah, J. H. & Kreminski, M. (2024): Homogenization Effects of Large Language Models on Human Creative Ideation. Proceedings of the 16th ACM Conference on Creativity & Cognition, 413–425. View the study