When generative artificial intelligence became widely available, schools got a new "ghost" to handle: texts that look perfect, answers that arrive in seconds, and students who experience knowledge as always just a keystroke away. Many teachers describe a double anxiety: How do we actually assess competence when a student can deliver an impressive draft in five minutes? And how do we safeguard mental health and psychological wellbeing when pace, pressure and expectations increase? In the book Teaching with AI (2024) José Antonio Bowen and C. Edward Watson offer a pragmatic way out of the panic. They argue that AI should not be treated as an enemy to be excluded, but as a tool to be understood, tamed and integrated into the school curriculum – in ways that genuinely strengthen learning, relationships and teacher resources. This article summarizes the book's main ideas and connects them to psychological perspectives on learning, child development and psychoeducation – including how this can be supported in digital offerings like balanced.ai.
Teaching with AI actually starts where many teachers already stand: Students submit texts that seem unusually mature, traditional tests lose value when AI can solve problems instantly, and attention is fragmented in a media landscape that rewards quick answers. The authors do not describe this as moral decline, but as a systemic challenge: Schools have long been built around scarcity of information, while students now live in abundance.
From a psychological perspective we can see the AI wave as a stress-test of three central factors in psychology in schools:
Bowen and Watson's core idea is that we will not return to "before AI." Therefore schools must move from bans and control to AI literacy as a new form of education – while also protecting students' learning and identity development.
The book is written by José Antonio Bowen (former president of Goucher College, known for Teaching Naked) and C. Edward Watson (vice president for digital innovation at AAC&U and head of an institute for AI, pedagogy and curriculum). Their perspective is clearly practice-oriented: Teachers do not need to become programmers, but they must understand the technology well enough to guide students through it.
A key claim is that AI is already integrated into everyday life: spellcheck, search, recommendation algorithms, spam filters and translation tools. Generative AI is "just" the next step – but with greater consequences for assessment, writing instruction and views of knowledge.
The book describes three foundational pillars for AI literacy. Translated to everyday school practice and psychological learning theory, they can be understood as:
Students must learn that AI is good at pattern recognition and language generation, but weaker on context, nuance and truth checking. The authors point out that AI can "hallucinate" – presenting errors as if they were facts.
Psychologically, this is about training epistemic vigilance (the ability to evaluate information critically). In child development this ability shifts from being trust-based (children believe authorities) to gradually more source-critical in adolescence. Schools must therefore provide structured training to distinguish the probable from the true.
Students must be able to spot generic phrasing, absence of concrete examples, "too smooth" argumentation and missing sources. This resembles classic source criticism, but with a new challenge: The text can be grammatically perfect and still wrong.
In adolescent psychology we know that social confirmation and "cognitive shortcuts" often guide decisions under time pressure. AI texts can give false security: "It sounds right." Learning to evaluate AI therefore also trains tolerance for doubt and delaying conclusions – closely linked to emotion regulation and stress management.
Bowen and Watson argue that AI works best as an idea partner, draft generator, "explanation machine" and producer of practice exercises – but poorly as a final authority, definitive answer or substitute for analysis.
This resembles a metacognitive skill: Students must learn where in the process the tool is useful. In school this means treating learning more as a process than as a product.
A strong contribution of Teaching with AI is the concrete strategies. The authors emphasize starting small, building habits and avoiding "overnight revolution."
Instead of making rules for students, the book suggests making them with them. This increases ownership and reduces resistance. Students often have more insight into actual use than adults, and can contribute to more realistic norms.
From a psychological perspective this strengthens autonomy (Self-Determination Theory): When students experience participation, the likelihood of intrinsic motivation and compliance increases. At the same time it can support attachment in the classroom: Rules become a shared project, not a control regime.
The authors emphasize that the quality of AI responses depends on the quality of the questions. Students must learn the difference between vague and specific prompts.
Practical example:
Pedagogically this is a form of "question competence" and academic writing skill. Psychologically it trains planning, precision and the ability to define a problem – core components of executive functions.
When assessment only measures the final product (essay, submission, test), AI becomes a shortcut. When assessment measures reflection and process, AI becomes a visible tool.
Practical step: ask students to submit a "work log" or "thinking diary" describing choices, sources, prompt, errors they found, and what they changed. This strengthens metacognition and makes learning harder to outsource.
The book recommends using AI to generate extra practice tasks, quiz questions, or a first draft of feedback that the teacher adapts. The goal is to free up teacher resources for the human aspects: relationships, guidance, conversations and tailored support.
Case (realistic school day): A language teacher can use AI to create five alternative writing prompts at three levels (basic, standard, advanced). The teacher uses the time saved for short guidance conversations with students struggling with structure or performance anxiety. The result can be both higher academic progress and better psychological wellbeing in the class.
The authors suggest regular exercises where students receive AI-generated text and must find errors, improve arguments and add sources. This makes AI's limitations concrete.
This is also a psychoeducational practice: Students learn about cognitive biases, authority bias and how "confident language" can mislead us. In addition, it can create a culture where mistakes are a learning tool, not a failure – a key for mental health in school.
One of the most interesting points in the book is how it suggests responding to AI-generated submissions: not primarily with punishment, but with curiosity and a learning dialogue. This does not mean everything is allowed; it means the goal is to strengthen judgment, responsibility and academic integrity.
A short formulation that summarizes the stance (and aligns with the book's message) can be expressed like this: "The goal is not to stop AI use, but to make it visible, reflective and academically relevant."
Psychologically this can reduce shame and defensiveness. Shame often increases hiding and avoidance, while openness increases learning. By asking: "What did you use AI for? Which parts are yours? Where are the weaknesses?", you train both ethical reflection and academic maturity.
Teaching with AI points out that traditional tests become less meaningful when AI can deliver rapid answers. The solution is not just "stricter control", but smarter assessment: oral defense, projects, portfolios, collaboration and reflection over time.
This aligns with research on learning and stress: When assessment feels like unpredictable control, performance anxiety increases. When assessment feels like a predictable process with clear criteria and opportunity for improvement, it can strengthen self-efficacy.
Concrete ideas (AI-robust assessments):
An important, sometimes underestimated point in the book is that AI can help teachers differentiate: create multiple levels of the same assignment, different explanations, and more practice. This can be extremely valuable in a classroom with wide variance.
Seen through the lens of child development this is about meeting the student in the zone of proximal development: tasks that are slightly challenging but achievable with support. AI can help produce "scaffolding materials" quickly (examples, step-by-step explanations, alternative formulations), while the teacher ensures quality and relationship.
Practical example: In math, AI can generate variants of equation problems with gradually increasing difficulty and step-by-step solutions. The teacher can then use the time to observe, ask questions, and support students who struggle with working memory or math anxiety.
Psychoeducation means giving people knowledge and language about psychological mechanisms – so they can understand themselves, regulate emotions and make better choices. In school, psychoeducation can cover stress, sleep, attention, social comparison, motivation and digital habits.
AI makes psychoeducation more relevant, not less, because:
Schools can meet this with psychoeducational micro-lessons tied to AI use, for example:
Digital tools can make psychoeducation more accessible, especially when teachers face limited time and capacity. An offering like balanced.ai (see bebalanced.ai) can function as a low-threshold resource for students and staff by supporting:
A practical way to link Teaching with AI to psychoeducation is to create an "AI week plan" in class where students both learn prompting and work on regulation skills:
Here balanced.ai can be used as support between lessons so that the insight becomes more than "a good lesson"—it becomes a habit.
Although Teaching with AI is solution-oriented, its content points to some obvious pitfalls:
Teaching with AI (2024) describes a school day many already recognize: AI is used, assessment is challenged, and attention is under pressure. Bowen and Watson's main contribution is to shift focus from fear to skills. They show that AI competency can be built around three pillars: understanding AI's strengths and limitations, being able to evaluate AI content critically, and integrating AI strategically into the learning process. They then offer concrete measures: create rules with students, teach prompting, assess process over product, use AI for practice and feedback, and train critical thinking via fact-checking.
Linked to psychology in schools, this is about more than technology. It's about motivation, self-efficacy, relationships, attachment and students' mental health. When AI is used wisely, it can provide more personalized instruction and free teacher time for what truly matters: safety, guidance and deep learning. And when psychoeducation is given space – both in the classroom and through supportive tools like balanced.ai – schools can meet the AI era with both academic sharpness and human warmth.