“AI liberates creativity in the mind, even in people who have no artistic talent”

INTERVIEW WITH STEPHAN ZIPPERLEN, COVESTRO

“AI liberates creativity in the mind, even in people who have no artistic talent”

How do you reach an agreement with the legal department on the low-risk, officially permitted use of generative AI in image and film production? And where does the use of AI avatars make sense? Stephan Zipperlen, Communications Excellence Catalyst at polymer specialist Covestro, answers these questions in an interview with Richard Tigges, Co-Lead of our Technology Cluster.

Stephan Zipperlen is an active member of the Technology Cluster of the CommTech working group. If you would like to exchange ideas with him and his cluster colleagues, just come along to the next online meeting. Just send an e-mail to richard.tigges@audi.de or cschmid@christofschmid.com and we will send you a team invitation!

Richard Tigges: Stephan, your title of Communications Excellence Catalyst sounds very cool. What’s behind it?

Stephan Zipperlen: I am responsible for communication on audiovisual communication and events – especially for our Executive Board and the Extended Leadership Team. This involves strategic planning, format development and implementation, the look and tone. I accompany these appearances and thus have a very interdisciplinary overview of our topics. Together with a very high personal affinity for technology and AI, I have slipped into an enablement role and have implemented an AI masterclass and pilot projects for transcribing and translating videos as well as the use of generative images in corporate communications in my team. As part of an internal campaign, we are currently preparing knowledge for all colleagues instead of relying on external trainers – I am the Product Owner for the General AI Offerings area. We are showing Covestro employees that the topic of AI is something we expressly want. The Board of Management is behind it so that we can go down the AI path or at least think about it, even if there are a few doubters in the hierarchy. AI is already an important building block today.

Richard Tigges: Other members of the CommTech working group report in our Technology cluster that it is still terribly complicated in many places. You have nevertheless ventured into the legally tricky topic of image generation?

Stephan Zipperlen: Since the results of Mid-Journey began to be rudimentarily useful a year and a half ago, we have sat down and analyzed the largely jurisdiction-free space. You can read through rules today, but the law hasn’t often been applied in court. So there are a few unspoken ambiguities. However, we didn’t want to be put off by this and didn’t want to wait until it was finally clarified, as this could take years. We found very cooperative discussion partners at IT Security and the legal department, carried out a classic risk assessment and initially concentrated on the monetary risks in copyright law. How do we avoid deliberate copyright infringement? A prompt to generate an image in the style of a well-known and contemporary artist. Or the desire to imitate a certain character from a movie. It is relatively obvious that these would be copyright infringements. As soon as architecture or art styles come into play, it gets very complicated. We have excluded such things in our guidelines from the outset.

Richard Tigges: Architecture too?

Stephan Zipperlen: If the picture is taken from a non-public perspective, for example from the balcony on the second floor of the house opposite, then the architect of the house shown has the rights to this picture. Nobody could have guessed that! So we completely omit concrete relationships or references to architecture. No building, no street. We have also excluded the use of image uploads to the AI, because otherwise we would have to ensure that we have the rights to the motifs for AI processing. We will be marking this in our image database soon so that we will soon be able to allow uploads. But that’s the beauty of such guidelines, that they can and may evolve.

Richard Tigges: What about negligent abuse?

Stephan Zipperlen: We have built a short reverse image search into our AI process so that we don’t just rely on the image generator, but also check to make sure that the motif is really unique. A small barrier against negligence.

Richard Tigges: Is it worth the effort for AI images?

Stephan Zipperlen: They certainly won’t replace photography in corporate communications, but I can really see it replacing the purchase of stock material. I believe that the topic of AI and images will change massively in the coming months and years.

Richard Tigges: Do you also work with AI-based video avatars?

Stephan Zipperlen: What’s exciting is that the language of videos can be changed and even the lip movement can be adapted for synchronization. The Ukrainian embassy has a Twitter channel with an AI avatar as a speaker. At Covestro, we don’t want to replace people, nor do we want to replace photography or videography. Even if it would be tempting for the CEO to save two hours for a video shoot and have his avatar speak instead, no. Where we see real added value is in safety issues, for example. In the chemical industry, we have to invest a lot of time and effort in process safety training. It makes a difference whether I read through a one-pager or watch a nice web-based training course that not only shows text but also an avatar in protective clothing. Spoken in the viewer’s native language, this provides a different approach to the content.

Richard Tigges: We are now very close to reality, but there is still a lot of artificiality in AI. What does it do to you to see artificial moving images or pictures? The visuals go straight to the heart. Do you sometimes miss the authenticity?

Stephan Zipperlen: I am highly fascinated by what is possible. I still see the value of authenticity, especially when it comes to people and the personal. One of the most exciting technological changes is the discovery of photography, which has managed to capture moments and use them to tell stories. Without having to have the expertise to master pencil, ink or oil drawings in depth. I see generative AI as a similarly important innovation, especially in the audiovisual sector. It creates something new and gives us the opportunity for creative expression. Some things still seem very artificial and strange, especially at the beginning. But when I look at the first literographies, history repeats itself. But the opportunity to express illustrative thoughts without having the talent to draw, digitally edit or illustrate is brilliant. Some of us are discovering an aesthetic, artistic side to ourselves. And of course we need to talk about authenticity, genuineness, watermarks or other clear labeling and, above all, create an awareness in society as a whole of what generative AI can and may do. Also important: how will we value artists who still put things on paper themselves in the future? At the same time, AI also frees up creativity in the minds of people who have no artistic talent – with generative AI, they can put their vision on paper. And that’s what impresses me most about this medium.