August 2023

Every newspaper writes about it and in private, this issue is a hot topic of discussion: AI chatbots like ChatGPT or Google Bard. Do they have the potential to change our communication, communities, and the future of work? This discussion is also taking place at Content5: How does this technology change our work – right now and in the future? How do we find the right balance between increasing efficiency on one side and meeting our standards for quality and data protection on the other? 

Far from only talking about it, we also put the different AI chatbots to the test by having them complete everyday tasks. For one of our customers, we are writing city profiles. Can the AI chatbot accomplish this? Our conclusion: Not just yet. The profile written by the AI chatbot contained false information and highlighted features that were not characteristic of the respective city at all. For other clients, a major part of our work consists of summarising news reports with a focus on issues that are relevant to that particular client. Can a chatbot summarise these reports for us? The deficits were obvious: Important information was left out, especially when there were multiple themes contained in the report. Even more difficulties arose when our output was based on multiple sources.

The Bottom Line

Efficiency-wise AI chatbots are, of course, without competition. However, at least for the time being, they still have numerous deficits which, together with the unclear data protection regulations, prevent us from using them. Due to our stringent data handling standards and requirements to continuously deliver high quality, the use of such chatbots will be limited and we will rely on them only in exceptions, e.g., to double-check results. We will continue to actively monitor developments and adjust our strategy when necessary.