The latest news about artificial intelligence is that Italy announced a temporary ban on the use of ChatGPT at the beginning of the month, while regulatory agencies in countries such as Germany, France, and Spain are closely monitoring the security issues of ChatGPT, especially in protecting minors and handling user data. Panic caused by ChatGPT is continuing to spread…
When it comes to the extent of AI’s subversion, international consulting firms generally believe that the three areas most affected are education, the working styles of ordinary white-collar workers, and the medical industry.
In fact, I believe that ideal education and ideal patient care both require a significant amount of attention with the ability to recognize and make judgments, taking into account real-time reactions and changing needs of students and patients. However, the current state of the industry is such that practices cannot provide the attention needed by the service recipients.
Today, I want to talk about the changes in deliberate practice that AI brings to the education industry.
Deliberate practice has been a popular form of education since the Industrial Revolution:
1. Classroom instruction is used to convey the content of the lesson.
2. Homework assigned by the teacher is used to complete exercises.
3. Testing and exams are used to provide feedback.
4. The comprehensive feedback is used to select and allocate resources. The transmission of knowledge, the development of skills, and the shaping of character are all largely accomplished through deliberate practice.
To a large extent, these deliberate practices are collective, including teaching methods and learning goals. Each child only receives different experiences and results, and rarely receives the individual attention they need.
In our daily lives, many people believe that “deliberate practice” is equivalent to “enduring hardship” and “cramming”, and it is an important part deeply ingrained in East Asian culture. Does this mean that “AI is here, and deliberate practice is useless”?
Differences and Mutual Reflection between AI and Human Brain
As we consider whether deliberate practice is necessary in the age of artificial intelligence, it’s important to objectively understand the different areas in which AI and the human brain excel, and to better understand what intelligence entails, including curiosity, creativity, and imagination.
This allows us to distinguish between the playing fields of humans and AI, identifying which arenas to avoid and which fields require cooperation with machines.
Differences between AI and Humans
AI is a creation of humans, a simulation of our own neural and cognitive processes, a projection of ourselves, but also a reflection of our own brains. Professor John Hopfield, a pioneer in AI, said that his motivation for studying AI was to understand how the human brain works.
Coming from a family of neuroscientists, his original interests were in psychology and physiology. His proposal was to simulate the countless neurons and their link strengths in the human brain from a biological perspective using large-scale neural networks (NeuralNet).
However, this approach was considered garbage by the mainstream reasoning school. Human neurons have a relatively short lifespan (<1 billion seconds) and low utilization, with many neurons unused and the information.
Understanding Intelligence and Curiosity
According to the theory of intelligence proposed by Sternberg, intelligence can be divided into three categories. When it comes to understanding intelligence, I like to use the concept of “whole brain intelligence.”
Despite its exceptional language capabilities, experience, and abilities, ChatGPT’s advantage lies only in analyzing the basic level of intelligence. It lacks perception and interaction capabilities with the environment and experience, making it a true narrow artificial intelligence. Although it can span many fields due to its language model, and provide valuable value to various fields, it still heavily relies on humans for adaptation and integration into scenarios and environments, rather than surpassing the singularity. In terms of creativity and practicality, there is still a long way to go.
Although some people consider this to be the starting point of general artificial intelligence, I believe it can still be classified as tool AI. It goes beyond the meaning of a search engine (information positioning) and is a better way of organizing, associating, combining, and mining information.
In fact, artificial intelligence is still observing and associating things, as shown in the first level of the following figure (from “The Book of Why: The New Science of Cause and Effect Paperback”). This association has no causal relationship.
Due to the lack of execution capabilities, or the inability to feedback the consequences of execution to observation and learning, it is impossible to reach the second level. The third level of imagination is more hypothetical.
As it is hypothetical, it has infinite possibilities. Therefore, in terms of intelligence, although artificial intelligence can be used by us, its scope is relatively limited.
D.E. Berlyn, a Canadian psychologist, classified curiosity into four categories after studying it:
1. Specific but only perceptual curiosity: such as curiosity about encyclopedic knowledge;
2. Specific but cognitive curiosity: such as scientific research-like curiosity;
3. Broad and only limited to perceptual curiosity: such as expectations of change;
4. Broad cognitive curiosity: I think it refers to readers who are not focused or generalists if they are all successful.
Machines or artificial intelligence are unlikely to have curiosity. But as humans, we should have more effective curiosity, which should be specific and cognitive.
Understanding Creativity and the Role of Hardship in Learning
Currently, creativity is generally understood to encompass three types of creation: combinational creativity (combining existing ideas A and B to generate new ideas), exploratory creativity (exploring and discovering in unknown areas), and transformational creativity (transforming the features of A into the creative type of B).
Artificial intelligence is more likely to achieve combinational and transformational creativity, while the possibility of AI achieving exploratory creativity is limited, requiring the establishment of framework and cognition by humans. This classification of creativity helps us understand the malleability of creativity in certain aspects, but it is based on a good cognitive foundation.
Neuroscience research has shown that the period before and after puberty is a golden time for children’s learning, as their brains are highly plastic and learning is fast. Plasticity is an important research direction in neuroscience, and research has shown that even in adulthood, the brain remains plastic and can adapt to changes caused by injuries or major life changes, often through repeated practice.
In terms of learning and work, Chinese and East Asian cultures have always regarded “eating bitter” as a symbol of effective learning or progress, even as a kind of faith. The concept of hardship in learning and growth is indeed reasonable.
Recent research has shown that confusion and overcoming confusion is the best time for effective learning, and some of the confusing and frustrating “bitterness” in learning can deepen memory. While the Confucian emphasis on hardship education is good, it was further expanded under the influence of the imperial examination system. This, along with the introduction of the Prussian education system (which trained disciplined laborers) and the subsequent Soviet education system (which rapidly trained professional talents), formed a solid “hardship education” fortress.
The Effectiveness of Deliberate Practice in Education
In China, the limited educational resources and job opportunities have led to a strong emphasis on hard work, with children often being pushed to practice relentlessly. One child in Tianjin was quoted as saying that when he opens his eyes, he doesn’t see the sun, but rather homework.
On the other side of the Pacific Ocean, many people in the United States and Canada blame the current teaching methods of “happy/exploratory mathematics” and “completely natural” language learning for not providing enough deliberate practice opportunities for children, resulting in a gap in basic skills.
A recent book, “The Intelligence Trap: Why Smart People Make Dumb Mistakes,” praised the “hard work” education in East Asia.
Deliberate practice is useful, but it should be done in a certain order (from basic skills to advanced) and with specific attention paid to certain aspects.
The differences between humans and machines that we discussed in the previous section, as well as my previous article on metacognition, provide good guidance on the focus of deliberate practice.
Here, I introduce a graph as a good guide: deliberate practice is most effective when both the “positive torture” and the establishment of clear knowledge links are considered. Without effective struggle, it is merely rote memorization, and if teachers only lecture fluently and students do not think or practice, they will not retain information and will not improve. Have any of you ever encountered such a teacher?
Effective deliberate practice is difficult to achieve in a collective educational setting because each individual has different goals and progress.
However, with the assistance of technology, deliberate practice can be gamified and made intelligent, while also taking into account the child’s emotions and physical needs, such as time scheduling and frequency of practice.
We must cultivate whole-brain children: those who can coordinate the left and right brain, integrate the brain’s fast and slow systems, and process their own emotions in their own way.
