On September 9, the Department of Computer Science and Technology at Tsinghua University released the Redbook "Artificial Intelligence General Education". It sets out the scholarly foundations and historical mission of AI general education (AIGE), and proposes a curriculum system and implementation plan centred on "cognitive autonomy". This series explains, in plainer language, the Redbook's overall structure, key concepts, main claims and central evidence.
As the seventh article in the series, this piece discusses how the AI general education course provides, on top of the existing curriculum, cognitive training and social value that are essential and irreplaceable, and on that basis argues for its independent standing in education.
The previous article discussed the definition of AI general education and how it becomes a new form of education. General education supplies cognitive autonomy as the unifying goal, while AI supplies the knowledge base and the training examples that allow this goal to be realized as a concrete process of learning. This article goes on to discuss the distinctive value of the AI general education course, and how that value supports its independent standing within the education system.
To discuss this question, we first need to be clear about the distinctive goal of the AI general education course in cognitive cultivation. Only by identifying this goal through comparison with existing courses can we go on to explain the mode of cultivation it adopts, the social function it performs, and the educational standing it ought to have.
1. The Goals of Cognitive Cultivation in Existing Courses and in the AI General Education Course
Courses such as mathematics, language, science, history and information technology already cultivate students' cognitive abilities from different angles. Through its particular knowledge and learning activities, each course forms its own focus of intellectual training. The Redbook selects five representative disciplines and compares them with AI general education, as summarized below.
| Discipline | Core intellectual training | Criteria of judgment |
|---|---|---|
| Mathematics | Necessary reasoning under formal constraints | Are the premises explicit? Is the reasoning valid? Does the conclusion follow necessarily from the premises? |
| Language and literature | Interpreting and expressing meaning in context | Is the understanding consistent with the context? Is the interpretation grounded in the text? Is the expression accurate and appropriate? |
| Physics | Causal explanation under empirical constraints | Is the explanation consistent with empirical facts? Is the mechanism plausible? Can the prediction be tested? |
| History | Contextual judgment from limited evidence | Are the sources reliable? Can the evidence be corroborated? Does the interpretation fit the historical context? |
| Computer science | Executable construction for complex problem solving | Can the problem be represented explicitly? Can the process be executed and the results tested? Can errors be located and corrected? |
| AIGE | Cognitive autonomy in intelligent environments | Are the sources credible and the evidence sufficient? Are the limits of the conclusion clear? Is trust in the model appropriately calibrated? Are judgment and responsibility clearly assigned? |
Source: Redbook, Table 8.1. The entries identify each discipline's central emphasis; they do not purport to define the full value of any discipline.
These comparisons show that existing courses already train human cognitive abilities systematically. The goal of cognitive training on which the AI general education course is anchored is a person's cognitive autonomy in intelligent environments: when machines can take part in selecting material, forming explanations and making judgments, can people actively make use of this capability while keeping command of their own understanding and action?
This goal is distinct from those of other courses. That distinctiveness is the basis for the further discussion of its value for cognitive cultivation, its social function and its independent standing.
2. Distinctive Value for Cognitive Cultivation
"Cognitive autonomy" is the disciplinary goal of the AI general education course. The difficulty is that cognitive autonomy is not something only an AI course can cultivate; other disciplines can do so too. What, then, is special about the AI general education course? How does it establish a distinctive value in cultivating the capacity for cognitive autonomy?
1. Educational goal: organizing a complete process of cultivation around cognitive autonomy
Intellectual training in each discipline usually unfolds around that discipline's own objects of study and educational goals, and "cognitive autonomy" is an important part of it. Students can learn to check premises in mathematical proof, come to understand the sources of evidence in studying history, and reflect on their own assumptions in scientific experiments. Such training is valuable and can cultivate the capacity for cognitive autonomy from different angles. However, it is spread across different courses and different tasks, each serving the main educational goal of its own discipline, and cannot form a continuous and complete process of cultivating cognitive autonomy.
The AI general education course advocated by the Redbook explicitly takes cognitive autonomy as its main educational goal. How knowledge is selected, how activities are organized, how learning progresses and how outcomes are assessed all need to be arranged around this goal. Whether students can form well-grounded understanding, whether they can test the conditions under which a judgment applies, whether they can notice their own trust and dependence, and whether they take responsibility for their actions: these constitute important criteria for examining whether the course has fulfilled its educational task.
This means that AI general education can organize training in the capacity for cognitive autonomy into an explicit and sustained process of cultivation. Cognitive autonomy is a basic capacity bearing on human development, and in an intelligent society it also faces the real pressure of being easily eroded. Education needs to train it in a sustained and systematic way, and cannot merely expect the relevant elements scattered across various courses to come together naturally and form this capacity of their own accord.
The AI general education course, therefore, first of all gives the mission of our time, "cultivating the capacity for cognitive autonomy", a clear home in the curriculum.
2. Mode of cultivation: learning AI knowledge and prototype-based training
A course with a clear goal also needs a mode of cultivation suited to that goal. The previous article explained that AI supplies both a body of knowledge and training examples. In the AI general education course, these two aspects appear respectively as knowledge learning and prototype-based training.
AI knowledge lays the scientific foundation for cognitive autonomy. In an intelligent society, people need to understand the systems they are drawing on: how such a system acquires its capabilities, how it produces results, why it may go wrong, and where its conditions of applicability and its limitations come from. These questions have to be explained through the concepts, principles and methods of AI. Only by understanding a system's basic mechanisms do people have grounds for judging which results deserve trust, which situations call for checking, and when a judgment should be made afresh.
Other disciplines can provide some training in the capacity for cognitive autonomy, but they cannot replace this body of specialized knowledge. A person who can reason rigorously does not thereby understand how models are trained and how they generalize; a person who can discern the stance of a text does not thereby understand how generative models and recommender systems work. AI knowledge allows general methods of judgment to enter concrete intelligent environments, and gives cognitive autonomy a knowledge base directly related to its real circumstances.
Prototype-based training, for its part, provides a concrete form of practice. AI systems can be turned into learning prototypes that retain the key mechanisms and are suitable for teaching. Learners can take apart their components, trace how they run, change the conditions and test the results, and around these objects study in a concentrated way how grounds are formed, how conclusions are produced, where errors occur, and what boundaries the system has.
The value of this training comes from the fact that AI systems themselves can be constructed, run and analysed. Students can put their own understanding into the form of data, rules, models or procedures, and then check that understanding by running the system. The cognitive predicaments encountered in an intelligent society thereby take on a form that can be observed and operated on.
Knowledge learning gives judgment its grounds; prototype-based training "hones" cognition in a concentrated way. Related knowledge and activities in other disciplines can support this process, but they cannot take the place of learning AI knowledge and analysing AI systems as prototypes.
3. Disciplinary character: coming to know one's own intelligence through AI
Artificial intelligence is the science of using computers to simulate intelligent human behaviour. This disciplinary character gives the AI general education course a further educational value: it can discuss human intelligence and artificial intelligence side by side, and, while dissecting the mechanisms of AI, help learners look back on their own processes of intelligence.
Analysing which samples a model acquires its capabilities from can prompt people to ask where their own experience comes from and whether it is representative; analysing which cues a model uses in making judgments can prompt people to examine what they themselves rely on when judging things; analysing how a model fails in new situations can prompt people to recognize the boundaries within which their own experience and conclusions apply. The study of machine intelligence thus becomes a concrete point of reference for understanding human cognitive processes.
The internal mechanisms of machines and of humans may differ, but both face such questions as how to form conclusions from limited experience, how to handle new situations, how to deal with uncertainty, and how to find errors and correct them. AI turns part of these processes into a computational form that can be analysed, allowing learners to think deeply about their own cognitive activity with the help of an external object.
This self-knowledge is an important foundation for building cognitive autonomy. People need gradually to become aware of how their own cognition is formed, what readily influences it, and under what conditions it may fail, before they can check their grounds and revise their judgments more consciously. The AI general education course therefore has a distinctive pedagogical feature: in coming to know AI, learners are also learning to know themselves. This gives cognitive autonomy a mode of training based on comparison between human and machine.
3. Distinctive Value in Social Function
The cultivation of cognitive autonomy also has a social significance that goes beyond individual development. Learning, work and public affairs in an intelligent society all need participants who can understand problems, form judgments and take responsibility. AI general education is addressed to all members of society, and its social function therefore has two interconnected levels: building a common cognitive foundation, and sustaining the cognitive subject in public life.
1. Building a common cognitive foundation for the intelligent society
General education has long carried one task: under conditions of differentiated knowledge and professional division of labour, to enable people of different backgrounds still to understand the world they live in together. Members of society may have different specialities, but they need a certain amount of common knowledge and common ways of knowing in order to communicate, cooperate and handle shared affairs. AI general education inherits this task and at the same time responds to changes in the way society operates.
AI is gradually entering the dissemination of information, the production of knowledge, organizational management and social decision-making, and is becoming an operating condition that people of different occupations and ages face in common. Only with a basic understanding of how AI works can one take part in the discussion of public issues. AI knowledge thereby acquires public significance, providing a common basis on which people of different professional backgrounds can exchange reasons, understand disagreements and take part in decisions.
2. Sustaining the cognitive subject in public life
Common knowledge enables people to communicate; public life also needs people who actually take part in judgment. Whether in organizational cooperation, public discussion or collective decision-making, participants need to understand reasons, evaluate evidence, make choices, and take responsibility for what they say and do. Here cognitive autonomy becomes one of the foundations of the subject in public life.
AI is entering such stages as obtaining material, organizing reasons and expressing opinions, making it the norm to take part in public life with the help of systems. At the same time, when these stages are carried out by systems on people's behalf over the long term, participants' opportunities to understand, verify and reflect for themselves may also diminish. Public discussion may remain very lively, while the judging subject behind it gradually weakens.
The special mission of AI general education lies precisely in explicitly bringing the cultivation of cognitive autonomy into an educational process addressed to all members of society, preserving citizens' standing as subjects so that they keep the capacity for independent thought and independent judgment when taking part in public life. This is a precondition for the stable functioning of a society.
The social value of the AI general education course therefore already goes beyond the popularization of tools or the training of specialists. It bears on whether the intelligent society can continue to sustain its members' independent thinking and sense of responsibility, and so provide the necessary cognitive foundation for the continued functioning of public life.
4. From Distinctive Value to Independent Standing
The analyses of cognitive cultivation and of social function together show that what the AI general education course takes on is a complete educational task.
From the standpoint of individual cognitive cultivation, the AI general education course carries the mission of systematically cultivating the capacity for cognitive autonomy. That it carries this mission is determined by the disciplinary character of AI, and other disciplines can hardly take its place. The AI general education course therefore has a structure of content and a task of cultivation that deserve to be kept whole.
From the standpoint of social value, the AI general education course carries the dual mission of building a common cognitive foundation and sustaining the standing of the cognitive subject in public life, which bears on the stability of society's basic operating mechanisms. AI general education therefore has a strongly public-education character, and needs to be supported uniformly through public expenditure and made compulsory.
The independent educational standing of the AI general education course is thereby established. It means that this course should have a clear place in the curriculum as a whole, with content, learning progression and assessment requirements suited to its goal, and with someone clearly responsible for delivering it.
Schools that have the conditions can offer it as a separate course, and in a transitional stage cross-course collaboration can be organized; whatever form is adopted, the completeness of the educational goal and of the process of cultivation needs to be ensured.
5. Conclusion
The independent educational standing of the AI general education course does not rest on how powerful or how useful AI is as a technology. The fundamental reason is that the complex cognitive environment of the intelligent society is eroding individuals' cognitive abilities, and thereby affecting the underlying foundations on which society operates. The goal of the AI general education course is precisely to preserve people's cognitive independence and sustain the stable functioning of society by cultivating the capacity for cognitive autonomy.
Enabling all citizens to receive AI general education on an equal footing is therefore the first test to be passed on the way to an intelligent society, and a responsibility that educators should shoulder.
For individual citizens, receiving AI general education is both a personal right and a social responsibility.