"We must not assume that AI can do everything"

Digital ethics expert Cornelia Diethelm warns against the "man-machine" comparison. Artificial intelligence can simply support and relieve us of certain tasks, but it is not a miracle weapon. The digital specialist also does not share the fear that AI will one day replace us.

Cornelia Diethelm lecturing at the HWZ University of Applied Sciences in Business Administration Zurich. (Picture: zVg.)

Artificial intelligence is not only omnipresent in newspaper articles, TV and online reports, AI has also arrived in the media and communications industry. Digital ethics expert Cornelia Diethelm will give a lecture on Forum for the Swiss print and media industry "Power Pur" in Schlieren on precisely this topic. In her presentation, Diethelm will dispel misconceptions and reveal the most successful strategies for integrating AI.

Today, we are seeing above all what could be technologically possible with artificial intelligence. There are more and more pilot tests and tools that are intended to simplify our lives with the help of smart technologies. However, there is a time lag before it becomes clear where these possibilities can be put to good use. The head of the CAS "Digital Ethics" course at the HWZ cites chatbots - computer programs that can simulate and process human conversations in written or verbal form - as an example. However, most of these technologies are not yet suitable for everyday use.

"Although the technology is there, it hasn't led to all companies using chatbots," says the expert. There is a feeling that AI has its digital fingers in every pie these days. According to Diethelm, however, this is not true because such systems are much more complicated than expected. "You have to look at each project individually to see whether and how well an AI implementation actually works." Such an introduction always involves "a lot of work" to ensure that it works properly at the end of the day.

First sound out the need

The 51-year-old cites the insurance company Helvetia and the Swiss Federal Supervisory Authority for Foundations as examples of corporations that actually use intelligent chatbots. "As with any implementation of AI, the need must always be sounded out." Just because there is a fascinating technology does not mean that the algorithms really make sense everywhere and will pay off at the end of the day.

If customers are more dissatisfied when they are served by a poor chatbot, then such an introduction makes no sense. It is also important to keep an eye on the costs. "The complexity of a project is often underestimated," says the Managing Director of the Center for Digital Responsibility. One example of this is self-driving cars. We are all still waiting for the technology to improve. "Autonomous driving is much more complex than we originally thought."

Diethelm is not suggesting that we should wait until the day after tomorrow to start an AI project. You simply have to think about what really makes sense. "A lot of thought needs to go into this in advance."

Calm down

Will much of what concerns AI never come true in the future, we want to know from the entrepreneur? Are certain circles and the media simply exaggerating here? "Certain headlines are intended to arouse interest and shake people up", so it's often just about clicks when it comes to artificial intelligence. Some people are also very alarmist. But they are a minority.

"We must not assume that AI can do everything." At the moment, generative AI dominates the scene, which only corresponds to a small part of what actually makes up our jobs and our lives. Artificial intelligence, which generates texts, images, videos and other data, is great. However, a professional local provider does not want subjects from the USA, but needs photographs from the region, which artificial intelligence cannot yet provide. "Just because it's possible to generate photos doesn't mean that everyone will be taking AI pictures in the future."

No "man-machine" comparisons, please!

Artificial intelligence is perhaps hyped so much by people and experts with a more technological understanding. However, AI is "only" about methods based on mathematics and statistics. And according to Diethelm, this is not identical to everyday life. "Our lives are not just about calculations and probabilities."
The "man-machine" comparison is being made more and more often, but AI really only deals with very specific and individual activities. Logically, artificial intelligence can calculate better than humans, "but that doesn't mean that the system is more intelligent".

The algorithm can relieve and support us: "But AI is not a miracle weapon and will not make us obsolete." A chatbot has to be set up by a human and then checked again and again to ensure the quality of the results is right. "Above all, we humans very often have to find solutions that chatbots can't cover," the Aargau native tells Werbewoche.

Man must look to the right

AI is not a system that runs by itself. "It still needs people or high-level expertise from people." The problem is that we currently focus too much on the machine and place it too firmly at the center.

The digital ethics expert describes artificial intelligence as a "means to an end", like a hammer that helps us hammer in a nail. However, even if we all had many practical devices at home today, humans would still have to do the household chores in the end.

Cornelia Diethelm advises that companies should start investing in the education and training of their employees when it comes to AI. A conceptual knowledge of what artificial intelligence is (and what it is not) should already be taught in schools. "If you understand how AI works, you can better judge where the algorithm can perform well and where it can't."

Internal training makes the company more independent. "This reduces the risk of a provider promising you the blue sky." To evaluate AI offerings, you only need a certain amount of basic knowledge. "AI really isn't rocket science!"

"Power Pur" - the forum for the Swiss print and media industry

Digital expert Cornelia Diethelm will on May 23 at the "Power Pur" event will give their presentation. Columnist Kurt Zimmermann, AI pioneer Chris Beyeler, diversity expert Esther-Mirjam de Boer and football referee Urs Meier will also be offering insights at the former NZZ printing plant, the "JED Events" in Schlieren. The overarching theme of the half-day event is "Reshaping Print & Media": specifically, it is about "Today's talents and tomorrow's technologies". The conference will focus on the shortage of skilled workers due to a lack of diversity and the challenges of artificial intelligence.

More articles on the topic