Final Thoughts
Rather than attempt to restrict the use of generative AI, our goal must be to teach students to use and surpass it. (Steipe, 2023a)
In the 18 months since the arrival of ChatGPT, we have seen a whirlwind of possibility, promise, failure, and growth in the use of GenAI tools for almost any task. Educators at all levels have had to grapple with changes to teaching and assessment methods, questioning the fundamental goals —and abilities—of education.
In navigating both our own use of LLM-based tools in STEM teaching, as well as our students’ tool use in their assignments, we are faced with a pedagogical evolution, if not a revolution. This book has described the history of AI in general and explained the development and training of LLM-based tools, as well as their limitations. We discussed biases inherent in the datasets that power these tools, underlining the urgent need for effective and ongoing mitigation efforts. We have presented some of the challenges of maintaining academic integrity, recognizing that while AI detectors and policies play a role, the solutions lie in open, constructive communication with students and innovative assessment strategies.
Educators are called to harness the strengths of LLMs in improving student engagement, enhancing learning materials, and providing career guidance, while also being vigilant of their limitations in areas like mathematical calculations and nuanced topic exploration. We have demonstrated the potential of ChatGPT as a facilitator for low-stakes learning activities (including developing prompt engineering skills), and as a partner in generating and refining content and assessments, offering a glimpse into a potential shift in the future of teaching and learning.
We have considered a variety of assessment formats, from oral exams to innovative uses of ChatGPT in generating formative feedback. These strategies underscore the shift towards assessments that reflect the changing nature of knowledge work in the digital age. Educators are encouraged to rethink the ways in which we evaluate understanding and skills, moving towards assessments that challenge students to synthesize, apply, and communicate knowledge in ways that machines cannot.
As we look forward, it is imperative that we continue to approach the integration of LLMs into education with a critical eye, embracing their potential to transform teaching and learning while remaining committed to the principles of equity, integrity, and critical thinking. But we need to put that critical eye on the nature of education itself: how will the power (for good, or for ill) of GenAI change the way we teach and assess? What is the value of education, based on societal needs?
The journey does not end here; it evolves with every update to the algorithms and every shift in the educational paradigms. We will continue to share our experiences, research, and insights into the effective use of LLMs in education, fostering an environment where technology enhances, rather than diminishes, the human aspects of teaching and learning.
