Contents
Fear of losing control, losing one’s job and fear of spreading fake news: How justified are these fears? The assessments of renowned experts differ widely.
Computer scientist and cognitive psychologist Geoffrey Hinton has been instrumental in the success of neural networks, the technology behind AI applications like ChatGPT. The native Brit has been researching this technology for 50 years, and now he’s sounding the alarm: The machines are on the way to overtaking us in terms of intelligence, and that’s not without consequences.
A look at history shows that the stupid can no longer control the more intelligent, according to the researcher. Hinton warns that machines will dominate humans. The fact that humans could lose control of super-intelligent systems seems realistic to philosopher Nik Bostrom.
Super Confined: Is AI a nerd?
Stuart Russell, Professor of Computer Science at the University of California and also an AI luminary, sees things very differently. He warns against overestimating the capabilities of AI systems. Russell sees the problems in the fact that AI systems are trained incorrectly today, that they stubbornly pursue a goal set by humans without worrying about the side effects.
Example social media: An AI was trained to compile the news feed to keep users on the platform for as long as possible, no matter how questionable the content is. If machines were intelligent, they wouldn’t make such mistakes.
The German science journalist Manuela Lenzen shows in a readable Book on natural and artificial intelligence impressive how big the differences between artificial and natural intelligence are – and how little people understand how intelligence comes about.
For the author, AI applications like ChatGPT are simply tools. It’s not the AI software that scares her, it’s the people who use these tools.
Man or machine: who is the problem?
AI-based text and image generators can produce and spread fake news on a large scale. Many renowned experts are therefore warning of a wave of fake news that is further dividing society.
The philosopher Joshua Habgood-Coote, on the other hand, argues that the dystopian warners misunderstand the problem: People are the problem and not the technology, according to the scientist, who conducts research at the University of Leeds.
Job loss: Forecasts are difficult
AI tools will change our lives, for example when writing texts, translating, illustrating or brainstorming – and thus also the world of work. Which jobs will disappear depends on how quickly the tools can overcome their many teething problems.
In the case of self-driving cars, this is slower than assumed ten years ago. Instead of the army of unemployed drivers predicted at the time, a serious shortage of drivers became a problem. Once again, the complexity of the task to be automated was underestimated and the potential of the machine overestimated.
Perhaps demanding tasks cannot be solved with today’s machine learning methods. Many researchers in the field of neuroinformatics, such as Benjamin Grewe from the University of Zurich, are convinced that physical experiences are essential for powerful AI systems.