Artificial intelligence (AI) is an extraordinary discipline. It has always aimed to determine whether and how it is possible to construct computers that come close to or even surpass the diverse capabilities of the brain – in other words, to technically decipher the central reason why humans are the life form that dominates this planet and are the only ones that can consciously influence its development. This has always made fictional and fact-based speculation in this area particularly appealing, and naturally not only for experts.
AI has long been an integral part of everyday life. It can be found in every search engine, every social network, every up-to-date recommendation algorithm, in translation, speech and image recognition services.
Countless people around the globe first learned how AI can now deal with written language when the American company OpenAI made its ChatGPT dialogue system widely available. What this comparatively new class of AI models is capable of is really very impressive – despite the many terrible mistakes it still makes.
Why are prominent AI players issuing warnings?
In recent weeks, two groups of prominent AI players have issued dramatic warnings about this key technology. In the more recent one, they even characterize AI as an "extinction risk" for humanity and compare the impending dangers with those of nuclear weapons or pandemics. In the past, the signatories include researchers such as Turing Award winners Geoffrey Hinton and Yoshua Bengio, but also entrepreneurs such as Deepmind founder Demis Hassabis, Microsoft co-founder Bill Gates and OpenAI boss Sam Altman.
It is quite justified to ask politicians to address the risks that can be associated with AI at the highest level. Incidentally, this is exactly what is already happening in America as well as in Europe and across the board, for example during the last G-7 meeting. AI will continue to influence social, cultural and commercial interactions, increasing and redistributing wealth and power. It can also be opportune to occasionally warn a little more vehemently in order to penetrate through the media background noise of this time at all – climate-sticking rioters also successfully use the end of the world as a dystopian figure of thought to make themselves heard.
However, anyone who issues such a far-reaching warning should at least explain in principle how the corresponding scenario could unfold in concrete terms. In the event of a nuclear war, this is self-explanatory, and in the case of a life-threatening contagious disease without an antidote, too. In the case of AI, it's not. Unfortunately, the current warners save themselves from going into this in more detail.
New types of weapons such as (partially) autonomous combat drones or swarms of combat drones are a realistic, dangerous AI application that has long been discussed. Misinformation in the form of texts, images or videos, which spread to an unprecedented extent and speed, is another risk that users, platform operators and legislators must learn to deal with. However, both phenomena are not fundamentally new. They also posed in the past as a result of technical progress, not least when the World Wide Web came into being.
Talk about risks early and seriously
What is absurd, on the other hand, is the now occasional fear that software or hardware equipped with its own (power) will and survival instinct could try to rise up against humans in some form. The AI systems that are so popular today are still far from human thinking skills. They can't really plan far ahead, they don't understand complex causal relationships, they don't have anything like "common sense", they don't even realize they exist, and they can't perceive themselves.
Some AI warners also have to be asked whether their words match their actions. Why did Altman – in contrast to the otherwise much-maligned Internet companies Google and Meta – make his voice AI easily accessible to the masses when such great damage threatens? No, AI won't wipe us out. It is a technology that offers incredible opportunities that can make life even more enjoyable in many areas. The risks must be discussed at an early stage, seriously and transparently. And wisely.