Data scientist lends insight to education in the era of AI writing  - SmartBrief

All Articles Education Insights Data scientist lends insight to education in the era of AI writing 

Data scientist lends insight to education in the era of AI writing 

The rise of AI writing means truthfulness, accuracy and integrity are even greater than before

6 min read

Insights

Woman's hands hold AI symbol balancing on a thread.Service of Ai Concept for article on AI writing

Francesco Carta fotografo/GettyImages

Insights is a SmartBrief Education Originals column that features perspectives from noted experts and leaders in education on the hot-button issues affecting schools and districts. All contributors are selected by the SmartBrief Education editorial team.

headshot of Eric Wang for article on AI writing
Wang

As far as I can tell, none of us were around for the invention of fire. But all of us today are living through an event of similar magnitude of importance: when the internet started talking to us.

I’m speaking, of course, about ChatGPT, the seemingly ubiquitous large language model chatbot that caught the world by storm late 2022 — just in time to wreak havoc on winter exams for educators around the world.  And it gets worse! It brought friends! New LLMs seemingly pop up every day: Bard, Claude and GPT-4 among others. 

These technologies, alone or built into a dizzying array of apps, are rapidly upending longstanding norms and expectations across a huge cross-section of fields and industries, but perhaps none more so than academic writing.

For the last 5,000 years, creating and understanding written language was the sole domain of humans, and suddenly it wasn’t anymore. On the face of it, this seems like a worst-case scenario.  However, when we peel back the doom and gloom, we see enormous opportunity to rethink and revitalize writing, both its role in the journeys of our students and how we will teach the next generation of writers.

Despite the enormous technological shifts brought by LLMs, I think writing is more important than ever. What’s more, I think the path forward on training the next generation of writers is a return to our roots — focusing on authenticity, integrity, human connection, trust, transparency and centering the student.

An act of defiance

Abstract ideas, fleeting thoughts and complex concepts do not naturally fall into a neat, orderly sequence of characters and words. They have to be wrestled into place through effort, thought and practice. Writing, then, is an act of defiance, one defined as much by the process as by the final product.

Through the writing process, students learn the habits of seeking and absorbing information, then prioritizing, organizing and communicating it to a specific audience. The advent of sophisticated AI-generated text does not change the importance of these skills; they are as important as ever. What will change, however, are the ways and mechanics of how we practice them, and the increasing importance of communication, feedback and clear expectation between educators and students.

The rise of LLMs and AI-generated text also puts a greater emphasis on the importance of truthfulness, accuracy and integrity. LLMs, for all their sophistication, still routinely and confidently make up facts, figures and sources. Students need to learn that even in courses that allow the use of LLMs, there is an expectation that they are responsible for ensuring that the submission is accurate, the sources are properly quoted and cited, and the interpretations are sound.

Wayfinding

We’ve seen in the immediate surge of popularity of ChatGPT that some schools and districts have moved to ban the use of AI-generated tools entirely. While decisions like these are understandable, I feel that in the long run they are both shortsighted and unsustainable. Not only are such bans unenforceable, they also deprive students of the opportunity to learn and understand the key skills around using AI effectively — skills they will need to compete for and thrive in the jobs of the future.

Educators will need to thoughtfully deliberate and collaborate on the role of LLMs in the classroom. While, ultimately there is no one-size-fits-all solution, it seems to me that the cornerstone of the LLM strategy is clear communication.  While some teachers may opt to permit AI-generated tools as long as they are properly cited, others might deem them off-limits or restrict them to draft creation. With such a diverse range of possibilities, how can you ensure your students feel confident and well-equipped to make the most of these emerging technologies?

  1. Co-create the LLM roadmap with students. Involve students in the complex conversations and decisions on the role of LLMs in the classroom.  Encouraging them to explore and articulate the benefits and risks of LLMs and incorporating their findings will foster a sense of ownership and responsibility in the guidelines.
  2. Emphasize proper citation. Discuss the importance of attribution, highlighting that AI-generated tools are themselves a source worthy of citation. This removes much of the stigma associated with using LLMs and prepares students for a future where human-AI collaborations are commonplace.
  3. Create space for dialogue. Encourage students to freely share their concerns, questions and ideas about incorporating LLMs into the curriculum and actively work to incorporate their feedback.
  4. Provide clear guidelines for LLMs. Incorporate unambiguous language around the expectations of LLM usage, when it’s appropriate and when it’s not, how it should be attributed and what consequences exist when trust is breached.
  5. Bring the community together. While some schools have parent communities that are technologically well-versed, other schools may find themselves as the experts on LLMs in their community.  Make sure parents, families and community members are brought into the conversation around LLMs as stakeholders, and reinforce the shared commitment to students’ academic success in the age of AI.
  6. Use tools for visibility. Use the latest generation of AI designed to find the statistical signature of LLM writing. These tools shouldn’t be used punitively, but rather to bring visibility that catalyzes discussion with students on the role of integrity and AI writing in the writing process.
  7. Trusting teaching instincts. No one knows their students better than their own teacher. Setting an initial style benchmark writing assignment at the beginning of a semester will give teachers a point of reference for a student’s style and level. Besides being a useful pedagogical tool, this benchmark will help identify AI writing.
  8. Up the assignment game. Instead of assigning general knowledge writing that can be answered through simple paraphrasing or synthesis, give assignments requiring critical thinking, close reading and deductive reasoning. Find ways to assess the process instead of only the final product. Collect notes and early drafts, and ask questions about the choices students made in assembling their final papers.

Igniting the human spirit 

For most of human history, fire was a dangerous, uncontrollable destructive force. But, taming fire was the first step our species took to where we are today. We stand at a similar crossroads now — and the choices we make will have enormous implications for generations to come. Let’s choose to embrace our journey forward, focusing on the things that define our humanity: our desire to communicate, to connect and to share.

Eric Wang, Ph.D., is the vice president of AI for Turnitin.

Opinions expressed by SmartBrief contributors are their own. 

_________________________

Subscribe to SmartBrief’s FREE email newsletter to see the latest hot topics on EdTech. It’s among SmartBrief’s more than 250 industry-focused newsletters.