Output Generation vs. Input Seeking: The Surprising Shift in AI Education That's Transforming Classrooms (A Must Read)

"By engaging AI as a thought partner, a sounding board for ideas, and a source of inspiration, teachers can unlock new levels of creativity and effectiveness in their teaching practices, leading to richer and more meaningful learning experiences for their students."

My jaw practically hit the desk when I stumbled across a study on teachers and Generative AI that casually mentioned 'backwards planning.' Perhaps it was the early hour or the lack of sufficient caffeine, but as someone who has spent countless hours crafting prompts with backward planning in mind – recognising its crucial role in AI literacy – this article struck a profound chord. It was a wave of validation and relief amid this crazy AI in Education whirlwind.

The often glacial pace at which educators are (embracing) this new frontier can be frustrating, yet it also fuels a fascinating exploration. Why the hesitation? Why the disparity in adoption? Some educators are diving headfirst into the AI World, eager to chart its depths and harness its transformative power. Others remain on the shore, cautiously observing, meticulously weighing the potential benefits against the perceived risks. This stark contrast ignites a burning curiosity within me: What factors drive these divergent responses? How can we seamlessly weave AI into existing pedagogical frameworks? And, perhaps most importantly, what will happen if schools don't adopt AI well?

The Backward Planning Paradigm: A Foundation for AI Integration

Backward planning, as articulated by Wiggins & McTighe (2005), "is the process of devising teaching plans that are well-aligned to learning objectives and goals in order to avoid aimless activities and “coverage” of topics" (p. 14). At its core, backward planning is about intentionally starting with the desired outcome – the specific knowledge, skills, and understanding you want your students to acquire – and then meticulously working backwards to design the learning experiences that will effectively guide them towards those goals.

This is precisely where generative AI can play a transformative role.

Unveiling the Dichotomy: Output Generation vs. Input Seeking

The recent case study of 24 US public school teachers, as detailed in "Backwards Planning with Generative AI: Case Study Evidence from US K12 Teachers" (Keppler, Sinchaisri, & Snyder, 2024), illuminated a fascinating dichotomy in how educators are currently leveraging generative AI within their backward planning workflows.

Some teachers focused primarily on utilising AI for output generation, prompting it to create specific teaching materials such as quizzes, worksheets, lesson plans, and even presentation slides. As Teacher 3, a high school math teacher, demonstrated during the observation period, AI can be a valuable tool for generating content within well-defined parameters: "Can you give me a puzzle where I have to find the next thing in a visual pattern... Give me an in-out table where the input is not a number but the output is a number... Make it a little more complicated" (Keppler et al., 2024, p. 17).

Teacher 5, a high school Spanish teacher, similarly leveraged AI's output generation capabilities: "Make 10 multiple choice questions for chapter 7 for the story Mi Proprio Auto... Make 10 multiple choice questions using the subjunctive for the story Mi Proprio Auto with an answer key" (Keppler et al., 2024, p. 17).

This approach to AI utilisation aligns with how generative AI is being employed in various workplace settings. Just as businesses are leveraging AI to automate routine tasks, upsell, generate reports, and create marketing materials, teachers can utilise AI to streamline the production of essential teaching resources, freeing up valuable time for other important tasks, such as providing individual student support or engaging in professional development.

However, the research revealed another, perhaps more nuanced, approach to AI integration. A subset of teachers in our study actively tapped into the potential of AI for input, seeking its insights and suggestions to refine their teaching strategies and gain a deeper understanding of their students' learning needs. These teachers, rather than simply asking AI to "make for me," engaged in a more iterative and collaborative process, prompting AI to "iterate with me" or "jumpstart for me."

Teacher 1, a middle school special education teacher, exemplified this approach during the observation period. She initiated a conversation with AI about the concept of adding positive and negative numbers, seeking its input on how to effectively teach this challenging concept to her students: "Can you explain how to add a negative number and a positive number?... Create real-world math problems within 100 that uses this concept… Add in multi-step word problems… Real-world examples using numbers within 20… What manipulatives besides a number line can I use?" (Keppler et al., 2024, p. 19).

Teacher 4, a high school ELA teacher, similarly engaged AI in a dialogue about the "great American novel," seeking its insights to inform her lesson planning: "Describe standards by which a “great American novel” is determined… Describe novels contemporary to Adventures of Huckleberry Finn that reflect similar social and cultural issues… Describe novels contemporary to The Great Gatsby that reflect similar social and cultural issues…" (Keppler et al., 2024, p. 19).

This approach to AI utilisation mirrors the strategic use of AI in business settings to gain insights into customer behavior, market trends, and operational efficiency. Just as businesses are leveraging AI to inform their decision-making processes, teachers CAN utilise AI to gain a deeper understanding of their student's learning processes and identify more effective teaching methodologies, leading to more impactful and personalised learning experiences for students.

The Productivity Paradox: Unveiling the Power of Input

The distinction observed between using AI for output generation versus input seeking proved to be a key differentiator in terms of perceived productivity gains. Teachers who actively sought AI's input on their teaching plans, those who engaged in a more iterative and collaborative process with the technology, reported experiencing significant improvements in both their workload and the quality of their work.

In the May 2024 survey, the researchers found that teachers who reported using generative AI for "iterate with me" purposes nearly twice as frequently as those who primarily used it for "make for me" purposes also reported significant productivity gains. These teachers felt they were able to accomplish more tasks in less time, and that the quality of their work had improved as a result of using AI. For example, as Keppler et al. (2024) noted, “The teachers in the first group—but not the second group—report productivity gains in terms of workload and work quality” (p. 3).

Teacher 13, a middle school science teacher, captured this sentiment well: "I think it saves me some time. But for the most part, I think it saves me stress instead... However, once I have a clear outline, then I feel like I’m able to move more efficiently and also feel more confident about it" (Keppler et al., 2024, p. 19).

This finding suggests that the true power of generative AI in education may lie not simply in its ability to automate tasks and generate content, but rather in its potential to enhance teachers' pedagogical thinking and decision-making processes. By engaging AI as a thought partner, a sounding board for ideas, and a source of inspiration, teachers can unlock new levels of creativity and effectiveness in their teaching practices, leading to richer and more meaningful learning experiences for their students.

Navigating the Complexities: Addressing Concerns and Ethical Considerations

The path towards integrating generative AI into education is not without its complexities and challenges. The research uncovered a spectrum of responses to AI, ranging from enthusiastic embracement to cautious skepticism and even outright aversion. These varied responses highlight the need for a thoughtful, nuanced, and ethically grounded approach to AI integration.

One common concern voiced by teachers was the potential for AI to facilitate cheating or erode students' writing skills. Teacher 2, a high school ELA teacher and a non-user of generative AI, expressed this concern succinctly: "it feels like cheating" (Keppler et al., 2024, p. 22). This concern is not unfounded, as the ease with which AI can generate high-quality text raises legitimate questions about academic integrity and the development of essential writing skills. As educators, we must carefully consider how to address this challenge and ensure that students are developing the critical thinking and writing skills they need to succeed in a world where AI is becoming increasingly prevalent.

Other teachers expressed concerns about the potential for AI to perpetuate biases or exacerbate existing inequalities in education. If the data used to train AI models reflects existing societal biases, then the outputs generated by these models may inadvertently reinforce those biases. This is a critical issue that requires careful attention to ensure that AI is used to promote equity and inclusivity in education, rather than perpetuating harmful stereotypes or disadvantaging certain groups of students.

Teacher 21, a high school social studies teacher and a non-user of AI, articulated a broader ethical concern about the very nature of generative AI: "I think ChatGPT kind of takes people’s work… broadly, kind of scraping the internet, to be trained on. And then it seems like large companies like OpenAI, broadly profit off said work… Conceptually, it’s bad" (Keppler et al., 2024, p. 23).

These ethical concerns highlight the importance of engaging in a broader societal dialogue about the responsible development and deployment of AI. We must carefully consider the potential implications of this powerful technology and establish ethical guidelines that ensure AI is used to benefit all members of society, particularly in sensitive areas like education where the stakes are high and the potential for unintended consequences is significant.

Towards a Balanced Approach: Embracing the Input, Guiding the Output

The research suggests that the most effective approach to integrating generative AI into education may lie in finding a balance between embracing its potential for input while carefully guiding its use for output generation. Teachers who reported the greatest productivity gains and the most positive experiences with AI were those who actively sought its input to enhance their pedagogical thinking and decision-making processes. They engaged AI as a thought partner, leveraging its insights to refine their teaching strategies and personalise learning experiences, as evidenced by the finding that "the teachers in the first group—but not the second group—report productivity gains in terms of workload and work quality" (Keppler et al., 2024, p. 3).

At the same time, these teachers also recognised the importance of maintaining their own pedagogical expertise and judgment. They did not simply accept AI's outputs at face value, but rather critically evaluated them, adapted them to their specific contexts, and ensured that they aligned with their own educational values and goals. For instance, Teacher 6, while using AI to generate a quiz, acknowledged the need for her own input: "I would absolutely need to insert illustrations for each of the questions, since a lot of the students at the age that I have are very visual with regards to their learning" (Keppler et al., 2024, p. 20).

This balanced approach reflects a broader understanding of the role of AI in the workplace. AI is not intended to replace human workers, but rather to augment their capabilities and empower them to achieve new levels of productivity and creativity. In the context of education, this means that AI can be a powerful tool to support teachers, but it should not be viewed as a substitute for their expertise, judgment, and passion for teaching. Rather, AI should be seen as a valuable partner in the educational journey, a tool that can help teachers create more engaging, personalised, and effective learning experiences for their students.

Charting the Course Forward: Questions for Reflection and Action

As we venture further into the age of generative AI in education, it is more important than ever that we engage in ongoing collaborative reflection and dialogue to ensure that this powerful technology is used to enhance, rather than diminish, the human element in teaching and learning.

Here are some key questions that we must grapple with as we chart the course forward.

1. How can we prepare future educators to effectively integrate generative AI into their teaching practices? What new skills and knowledge will teachers need to thrive in an AI-powered classroom? What training do teachers REALLY need? Prompt engineering? The best magic app? How to analyse your own processes? It's quite challenging to know which approach will work best. How can we equip teachers with the pedagogical and technical expertise they need to leverage AI effectively and responsibly?
2. What role should students play in shaping the development and deployment of AI in education? How can we ensure that students' voices are heard and their needs are considered as we integrate AI into the learning environment? How can we solve the problem of students (already) using AI vs those who are not using it? How can we best train students? How can we empower students to be active participants in the design and implementation of AI-powered educational tools and systems?
3. How can we foster a culture of continuous learning and adaptation within the education sector to keep pace with the rapid advancements in AI? How can we create systems and structures that support teachers in their ongoing professional development and their exploration of new technologies? Will we be taking away from key requirements such as literacy or numeracy training? Will Science and the Arts get cut from the team again? How can we ensure that the education system is agile and responsive to the ever-evolving landscape of AI?

We must engage in a collective effort to address these challenges and seize the opportunities presented by generative AI.

Let's get to work

Phil

References

Keppler, S., Sinchaisri, W. P., & Snyder, C. (2024). Backwards planning with generative AI: Case study evidence from US K12 teachers.

Wiggins, G. P., & McTighe, J. (2005). Understanding by design (Expanded 2nd ed.). Association for Supervision and Curriculum Development.