The age of artificial intelligence, besides promoting industrial and technological transformation, has also brought anxieties about the future of humanity. This anxiety stems from the fear that "most human jobs will be replaced by machines." In my opinion, however, the true impact of AI will likely be on jobs at the top of the value chain, involving values or requiring extremely high intellectual skills. So, in the age of AI, how should humans and robots divide labor and collaborate?
The smarter robots become, the more humans need to understand them.
In the age of artificial intelligence, one essential skill for humans is the ability to coexist with AI. Whether we like it or not, this era has arrived. In the next five to ten, or even twenty years, we will see its widespread application across various fields. The ability to coexist with AI, much like our ability to coexist with the mobile internet today, will become a necessity in everyone's life.
The more advanced technology becomes, the more humanity needs to understand it; the smarter robots become, the more we need to understand them, and to transcend their limitations to explore the boundaries of technology. So how can we develop this scientific mindset and keep pace with technological advancements? We need to return to the basics and understand the essence of science and the essence of artificial intelligence.
Artificial intelligence, at its core, is an information processing algorithm. The most promising algorithm currently is called deep learning, which mimics the neural networks of the human cerebral cortex. However, while the human cerebral cortex has six layers of neural networks, current AI programs can have over a hundred, hence the name deep learning. But no matter how complex the algorithm itself is, we can still understand its essence: using a deep network to process data and obtain the correct answer.
When we use this fundamental thinking to understand artificial intelligence, we can easily know what it can and cannot do in the future, starting with its ability to interact with people. While a deep neural network like AI can process massive amounts of data and information, it also has significant limitations. For example, it lacks the ability for self-reflection, the ability to truly understand common sense about the world, and the ability to comprehend the thoughts and feelings of others.
While current artificial intelligence can converse with humans, it can only process specific data within specific domains and lacks the vast background knowledge and common-sense understanding of the world that humans possess. It can recognize large amounts of data and images, but truly understanding a situation and the thoughts of each person within that situation is extremely difficult. Furthermore, artificial intelligence lacks many of the physiological functions of humans, including emotions, empathy neurons, and mirror neurons.
Artificial intelligence algorithms lack self-awareness, self-reflection, and self-consciousness, making it difficult for them to understand themselves. This self-understanding makes it difficult to understand others, because we often empathize with others by projecting our own experiences onto them—a capability artificial intelligence lacks.
Due to these limitations, in the future, artificial intelligence will excel in repetitive tasks with clearly defined rules, while humans will be responsible for communication and understanding others' intentions through extensive interaction. Ultimately, machines serve humans; the higher humans' position, the greater the assistance robots can provide.
