AI cannot replace leadership.

“The algorithms create nothing new – they draw only from what humans have previously created. And not everything we humans have produced over time is good – we’ve also created a lot of rubbish.”

Helena Nordman-Knutson calls for policies and personal responsibility.
This is written by Helena Nordman-Knutson, Senior Advisor at Safir Communication, in a guest column.

We have now entered the AI era, where artificial intelligence is driving technological development, including in the field of communication and content creation. Generative AI is expected to become as significant a force in innovation as the steam engine or the internet.

Many already use chatbot applications as writing tools – ChatGPT and Jasper AI are just two examples. These tools use advanced algorithms to generate text, videos, and images. They can optimize word choice, analyze language patterns, and translate content. You can even ask ChatGPT to rewrite a text in the style of a specific author.

These applications pull data from across the entire internet – including inaccurate content and copyrighted material. Because the writing tools pull from all kinds of sources, there is a real risk of bias. For example, research has shown that these tools describe men and women using different language, simply because that’s how men and women have historically been portrayed (Scientific American, ChatGPT Replicates Gender Bias in Recommendation Letters, Nov. 22, 2023).

With this in mind, professionals must be aware of the risks and apply a human filter when using AI-generated content. Demographic, discriminatory, and cultural challenges, misinformation, and clichés are just a few areas that require careful oversight. All content produced by these tools must be reviewed and quality-checked by a human before being published.

Publicly listed companies should adopt clear AI policies to ensure this is done. After all, the algorithms don’t create anything new – they merely recycle what humans have already produced. And not everything created by humans is of high quality – there’s plenty of junk too.

Many large Swedish listed companies now use English as their corporate language and draft texts in English first. Each language has its own nuances, and when English texts are translated into Swedish using AI tools, the results can be flawed. Direct translations rarely work. A word might not even exist in Swedish, and a short, concise English sentence can become twice as long in Swedish.

Buzzwords like “big picture,” “bottom line,” “flywheel effect,” “pain points,” “thought leader,” “upskilling,” “headwinds,” “leverage,” “resilience,” “retargeting,” and “future proof” are frequently used in English and then awkwardly translated into Swedish. These phrases are not always easy to render accurately, and translation demands precision.

Advanced algorithms that translate context rather than word-for-word may help, but they’re not enough. The translation process also requires a human filter – someone familiar with the company’s tone of voice and terminology. A person who can correct language mistakes and ensure that the texts are clear and coherent.

Various AI tools can also enhance efficiency in Investor Relations (IR). AI is becoming an active part of the IR toolkit. ChatGPT, for instance, can be used to draft frameworks for quarterly reports and press releases. AI can also help craft targeted investor presentations, generate insightful graphs, and tailor messaging for specific investor segments.

AI can transcribe earnings calls or serve as a chatbot for IR communication. There are many useful AI applications for daily IR activities – as long as they are used with an understanding of what generative AI is and the awareness that the output will only ever be as good as the input.

CEO letters are often written in a collective process. It usually starts with a draft from the CEO, CFO, or someone in IR/communications – or a consultant. The text is then edited, reviewed, and sent back and forth between stakeholders, before eventually going to the board and returning for final adjustments before the report is published.

This cycle can result in CEO letters that are so processed and sanitized that they may as well have been generated by AI.

I still believe that authentic, clear, and accessible CEO letters add real value – even if they take time to produce. Good examples of well-written, communicative, and personal texts include Johan Forssell’s CEO letter in Investor’s year-end report for 2023, or Johan Torgeby’s letter in SEB’s most recent quarterly report.

We are storytellers. Homo narrans. We’ve been telling stories since we gathered around the campfire, and I believe that the ability to communicate is a fundamental requirement for good leadership.