Wann man KI nicht in der Unternehmenskommunikation einsetzen sollte
Media Work

When not to use AI in Business Communication

Business communication is increasingly using artificial Intelligence in various formats. While chatbots have changed how consumers connect with businesses, targeted innovative marketing campaigns use AI algorithms to create customer groups who will most likely engage with the ad. As more people continue to use voice search, content is being adapted to ensure a seamless transition to voice assistants. PR has also leveraged machine learning and automation to gain better customer insights, conduct sentiment analysis, and ensure that various processes are streamlined and automated.

In theory, PR is the perfect setup for AI.

Not only can it easily accomplish vast amounts of data processing, but with Natural Language Processing, machines can even learn to mimic humans, and communicators can automate the tedious and repetitive tasks of PR, such as tracking email responses, publishing press releases, or even social monitoring.

However, AI as a technology also has limitations that must be at the forefront while making campaign and communication decisions. There have been several instances where it has resulted in PR nightmares for corporations. One of the most well-known examples is Microsoft’s Twitter chatbot Tay. Set up as an AI persona targeting the 18-24 demographic, the chatbot, whose purpose was to emulate the language of millennial women, began to spout Nazi sentiments due to interaction with toxic users and was taken offline after just 16 hours.

Another example with more severe consequences was IBM Watson’s computing system for oncology. Supposedly designed to help physicians make decisions, IBM Watson was widely criticised because it gave incorrect and unsafe recommendations that endangered patients’ lives.

Reputational risks while using AI in business communication can potentially harm brands as algorithms that make decisions are not transparent about the processes they use to reach those conclusions. Often, they use probabilities that can be flawed because of incorrect or biased data. Also, AI’s inherent racial and gender bias is still a work in progress, mainly as it uses patterns and insights but amplifies the biases from which it learns. It has already led to undesirable results and negative PR for corporations like Amazon. Another significant concern in AI is ethical and privacy issues related to data mining and breaches.

Deploying AI can be a business risk to your brand. In all the above examples, the PR team managed to control the narrative and avert plummeting stocks. Regarding business communication, humans must work in hybrid models to ensure that errors are minimised, smooth processes are in place, and machines are adequately trained to perform specific tasks.

For instance, while Danske Bank uses AI to monitor potential money laundering, it still relies on human judgment to examine the flagged transactions. Take the case of Unilever, which uses artificial intelligence to recruit and train employees; those who pass the initial rounds are invited for in-person interviews, where humans make the hiring decisions.

Regarding marketing and communication, it’s important to remember that they are about human connections and the relationships between brands and consumers or brands and journalists. For PR to be effective, it needs a human touch because machines cannot entirely replicate emotions. Journalists are more likely to respond to authentic sources instead of AI tech because the latter might make them feel manipulated. While automation makes smart replies possible, keywords can’t replace human intuition or contextual creativity.

The world has become digital at every step, and while automation makes life easier, AI is not a one-size-fits-all solution. It should not be considered a replacement for people but a judicious collaboration between humans and human-made systems, calling for a cautious investment that measures and understands the various risks involved.