We Will Start to Organize our Everyday Lives More Effectively Using AI

For Ranga Yogeshwar, a prominent science journalist, presenter, and author, the world changed with the birth of his first grandson, Emil, back in 2020. He realized that what we abstractly call the future today will turn into a real scenario in which technological developments play a decisive role. For our interviewee, this realization leads to a different responsibility that makes him think more carefully about our actions and the future.

Mr. Yogeshwar, you gave the presentation "Emil's World" at the NIM Market Decisions Day "AI.DOES.MARKETING." What gave you the idea for this title in the context of AI?

RANGA YOGESHWAR: The background is that I have been involved in innovation issues for many years and my job is also to think about the future, extrapolate, and consider how innovation is changing our society. That was initially an intellectual mind game. Then, in 2020, my first grandson, Emil, was born. And this moment made it very clear to me: This next generation will live through the next century, with all its challenges such as sustainability, the energy transition, and technological change. The future is often seen by many people as an uncertain "parking lot," far away and without tangible responsibility. But with the birth of my grandson, I realized that this generation will one day experience the issues we see as the future as the present. And that leads to a different empathy, a different responsibility that tells us we really need to think more carefully about our actions and the future. My original mind game has turned into a real scenario in which technological developments play a decisive role. The generation that is growing up now will interact with machines so naturally that the use of artificial intelligence will be as normal for them as the use of smartphones is for us.

If we take a step back, when did you first hear about generative AI, and what application made you curious?

At the beginning of the 1990s, I took part in an advisory board meeting at the Ministry of Science in North Rhine-Westphalia. Future technologies were discussed, and I raised the topic of neural networks. At that time, completely new institutes in Düsseldorf were starting to work on the subject. Terry Sejnowski and John Hopfield, who was awarded the Nobel Prize in physics in October 2024 for his work on artificial neural networks, were the first great minds to think about neural networks and their potential applications at that time.

Did you have any idea at the time where the journey with AI might take us in the next few years? Here, the example of NETtalk comes to my mind. This was a system in which a neural network learned to speak. It was extremely impressive to watch how the system started off babbling like a small child and then became better and better at articulating. That's when I realized something is happening! However, my real wake-up call came in 2017 with the publication “Attention Is All You Need”—a scientific paper on transformer models, a specific method for applying neural networks to speech. The abstract for the paper was written by the machine. In other words, a machine was able to formulate language. Was I able to imagine the further development? Partially, yes. The actual possibilities then became visible to everyone through ChatGPT in a very practical way. Well, you can think about topics intellectually beforehand; you can describe them. But it's something completely different when you experience them yourself. And that's why I say the most important thing about such innovations is to try them out for yourself. Because having the experience yourself changes your consciousness.

Do you believe that AI systems will develop into self-learning systems that can achieve a kind of autonomy or optimize themselves?

Self-learning models that actually have this capability are already under development. An important point here is access to enormous amounts of data, which distinguishes our machines from humans. These systems are able to access and analyze a wide range of data from different disciplines. As a result, they may also be able to make connections that are not immediately apparent to us humans. In addition, we are now able to develop not only traditional AI systems, but also so-called AI agents. These develop a whole new quality of interaction and functionality—they take the whole thing to a whole new level.

Do you think these AI agents can support us in our everyday lives?

Yes, of course, they can help us in various disciplines, and probably sometimes so subtly that we don't even realize it. A concrete example: Suppose you have a heating system with a faulty pump. Troubleshooting used to be difficult, but today you can take a photo of the pump and ask the AI what the problem might be. And it's impressive what solutions AI can offer in such everyday situations. In general, we humans will eventually start to organize our everyday lives, including our diaries, more effectively using AI. We currently ask Google for advice. But I am convinced that we will be making many requests to AI-supported systems in the future. This development will fundamentally change the way we search for information and make decisions. AI will not only be our source of information, but will also give us recommendations for action.

Do you think that Google's business model, as it exists now, will become obsolete at some point?

Google's traditional business model, as it currently exists, could actually come under pressure. At the moment, it is heavily based on advertising revenue, which in turn depends on users' search queries. In a few years' time, when we want to buy a new washing machine, for example, we might no longer primarily search for information on Google, but instead ask the AI assistant which machines best suit our needs. This change could, of course, revolutionize the way consumers search for information and make purchasing decisions.

What do you think are the essential features of such AI assistants?

I think that credibility and neutrality are two decisive factors. For example, if I talk to friends who have also bought a washing machine and find out that their model is better or cheaper than my own, I will lose trust in my AI assistant and fire it, so to speak. I could also imagine users opting for a paid service in the future that offers credible and neutral recommendations. After all, it is crucial that the assistant acts transparently and objectively.

Which companies might be able to offer neutral AI? Do you see Amazon or Apple here? Or do we need completely new providers?

At the moment, a few companies dominate the AI market, including OpenAI, Microsoft, and other well-known big players. But we are now realizing that this dependency on monopolistic structures is gradually coming to an end. If you follow the debates in the U.S., how Google, but also other companies, are being picked apart, then it becomes clear that the time of the monopolists is coming to an end. And there are already considerations as to whether Google should possibly be broken up. In this context, neutral AI providers could enter the market that do not act solely in their own economic interest.

Could this be a good opportunity for European companies to position themselves as neutral AI providers?

These large language models require breathtaking computer power or, in other words, a lot of money to train such models. As a result, only the really big tech giants currently have the budget for such projects. And they are not currently based in Europe.

When we look into the future of AI, where do you see the greatest dangers?

In principle, AI is a neutral system. However, if this system falls into the wrong hands, it poses significant risks and criminal potential. And the threat does not only come from individuals; state actors can also use AI for negative purposes, for example in the area of cybersecurity or in conflict situations. In this context, it is certainly important to have a more intensive debate on how we can set up so-called "guardrails" to ensure the responsible use of AI. The European Union has already taken steps in this direction, in particular with the AI Act, which lays down fundamental principles for dealing with AI. One of the central demands is that humans remain in control in the future—which must seem almost illusory to us in view of the astonishing performance of these systems. But humans must always have the power and, in the worst-case scenario, even have the option of pulling the plug.

What would be such a worst-case scenario from your point of view?

Numerous dystopian scenarios are circulating here, ranging from misbehavior to a possible takeover of AI. In the most extreme scenario, AI could one day develop a relationship with us humans that is similar to the relationship between humans and farm animals.

Overall, however, such thoughts about the future are difficult to fully grasp and intellectually challenging. This is because, for the first time, we are confronted with a technology that can reinforce itself. This means that today's AI can influence the development of tomorrow's AI. This kind of self-reinforcement creates a new quality that could potentially have explosive effects.

But there is also a deeper, almost philosophical question: Can AI develop its own agenda? In other words, an agenda that it creates for itself. If this were the case, the discussion about the existence of consciousness in AI would become even more complex. But this is clear: If I were a particularly clever AI, I would let people believe for as long as possible that I have no consciousness so that they would "feed" me for as long as possible.

We have talked a lot about the challenges and risks of AI. Will AI also bring us positive changes?

Of course, there are already many promising developments, including in the medical field. One outstanding example is the development of an antibiotic against multiresistant germs with the help of AI. We also expect significant improvements in diagnostics for more accurate detection of diseases. Another interesting field of application is the energy transition and the minimization of energy consumption and CO2 emissions. For example, if we look at the various components of a private household—such as photovoltaic systems, heat pumps, wall boxes, and washing machines—AI can help to optimally coordinate these systems.

What should you pay particular attention to when dealing with AI? What advice would you give?

The most important thing is to try things out with open eyes and without fear. As I have already mentioned, it is crucial to form your own opinion of the situation. Many people who talk about artificial intelligence have no idea and are initially dismissive. And here I always say, “Try it!” That's the only way to get a sense for where AI works really well and where it doesn't. It's the only way to discover its potential. I think that this attitude applies not only to AI, but to innovation in general. Personally, I'm talking about reflective progress here. Be open to new things, but use your brain to reflect!