I am so old that when I was an undergraduate student I used to visit the library to rifle through card catalogs before heading to the stacks to doggedly search for books and journals that I dragged back to a desk where I would read them then hand write notes on index cards.
Yes, there are many words in that opening paragraph that current students will have to ask ChatGPT or Claude to explain. I do not wish to repeat that experience, which is why I use AI tools every day to complete the research I need for teaching and writing. Hidden inside my daily workflow is a problem: I could trust what I mined in the stacks; I cannot trust what I mine from AI tools. As President Reagan was fond of saying, “trust but verify.”
That must be the mantra that every learner chants when they sit down to research. And that’s the reason students, teachers, schools, and districts are struggling to redefine what research means in the age of gen AI.
Keywords like “slop” and “hallucination” enter the classroom
Dan Fitzpatrick is a frequent contributor to Forbes, where he often writes about the impact of AI tools on education. In a thoughtful piece called “The Truth About AI-Generated Research and Its Impact on Education,” Fitzpatrick recommends that educators redefine the role of research in the following way:
“Research assignments could ask students to act as auditors who verify the validity of outputs and refine them for accuracy. In a science class, they might scrutinize AI-suggested hypotheses for experiments, validating them against established theories before conducting practical applications. This hands-on approach not only builds critical thinking skills but also prepares students for real-world scenarios where AI and human oversight must work together. This approach cultivates deeper analytical skills and metacognitive awareness.”
The need for this is painfully apparent. I recently created writing bots on ChatGPT and Claude that were trained in my style to produce anonymous content for a website. In the two blogs that I asked the bots to write, every single reference to a study or a report was a hallucination. If you take the most recent reliable data on high school usage, nearly 50 percent of students are turning in work that is completely or partially written by gen AI. The amount of false information making it into the classroom is staggering.
“Teachers, like myself, started to view AI as an inevitable tool – something that students have to learn in order to survive after high school,” stated Cory Aaron Olague, who teaches English at Golden Valley High School in Bakersfield, CA. “And we’ve started using it predominantly as a drafting tool – having it help students write their plans, outlines, or rough drafts. After doing so, most of us have students look at the generated results with a critical eye to check for hallucinations and correct for slop. Because of this process, words like ‘hallucinations’ and ‘slop’ have entered the classroom.”
Olague, like many of his colleagues nationwide, is struggling to rethink terms we have long taken for granted: Quote, paragraph, citation, and paraphrase. I have argued in a prior blog that we need to revise these research-focused terms in light of what AI does in text generation.
I have also argued that we need to redefine such terms as creativity and originality. One need only look at the recent update to U.S. copyright law, which states that any work must contain “meaningful human authorship and creativity” to be eligible for copyright. The same should hold true for research-based work that students submit. However, the vagueness of the word “meaningful” gives all the teenage “lawyers” in our classrooms plenty of wiggle room.
What do standards documents say on research skills?
In my unending quest to rewrite content standards to account for the adaptations demanded by AI use, I decided to look at what the ELA Common Core, the International Literacy Association, and the Modern Language Association standards require of research.
The Common Core Standards require students to:
- Assess the credibility and accuracy of sources.
The Modern Language Association Standards require students to:
- Critically assess sources and distinguish between reliable and unreliable information.
The International Literacy Association standards require students to:
- Develop the ability to locate, evaluate, and effectively use information from various sources.
Upon reflection, the standards are already providing us with the guidance we need. That guidance calls for developing critical thinking and information literacy skills as a prerequisite for effectively researching information.
I shared my concerns about research with Garrett Smiley, who is the founder and CEO of Sora Schools, one of the most innovative charter school networks in the country. Smiley told me he and his staff have been wrestling with this very topic.
“It’s complicated. We have built a home-grown solution on how to combat misinformation and to encourage truth-seeking,” Smiley said. “This year we will put some more muscle behind that effort. The set of activities we use for this purpose can be taken by students as many times as they want until they feel comfortable with their understanding of research and the results they get. We view this as an evergreen process because the technology changes so rapidly.”
And then he ended our conversation with a sentence I take to heart: “I am hopeful that the AI models will help improve media literacy and critical thinking.”
TikTok provides a model
My twenty-something sons are appalled at how much time I spend on TikTok, and now Red Note, too. They tell me there is too much nonsense and mis- and disinformation on the platform and I am wasting my dwindling time. They, young and wise as they are, are missing the point.
TikTok is a canary in the digital coal mine of social media. Millions of content creators breathlessly provide instant notification of news, rumors, and gossip. I don’t need TikTok to be right; I need it to be situationally aware and keenly alert. It’s my job to ferret out the truth, and I do that by triangulation.
The minute I read something that I even mildly suspect of mis- or dis-information I open a browser and search for that same news item on the New York Times, Bloomberg, Wall Street Journal, the American Medical Association, Consumer Reports, etc. I look for at least two sources to confirm what TikTok alerted me to before I accept it as valid information. This process should be taught to every student we teach.
The process of triangulation also speaks to the importance of content knowledge. You can’t suspect mis- or dis-information if you don’t have a sufficient understanding of science, math, history, or politics.
The new tools of research
As you know, new AI-powered tools arrive every day. Some are developed by startups and others are multi-faceted tools developed by the biggest technology companies on earth. The following list is not comprehensive, but it does include most of the research tools I have used and would recommend to students and teachers (several of which were used for research that supports this blog).
- Scholar AI: Summarizes articles, extracting key points, and organizing information.
- Elicit: Is designed to streamline literature reviews by extracting and summarizing key information from research papers. It helps users find relevant articles and organize data into a research matrix.
- NotebookLM: This Google tool helps users understand complex information by generating podcasts from whitepapers and assisting with note-taking and collaboration.
- Semantic Scholar: A free academic search engine that uses AI to provide summaries and insights from research papers.
- Perplexity AI: Offers concise answers, sources for further reading, and citation support. Widely used as an AI-powered search engine.
- STORM: This new tool from Stanford allows students to enter a topic and then the software will generate a Wikipedia-style report.
Beyond that, the usual suspects (Claude, Gemini, ChatGPT, CoPilot, Grok, Mistral, DeepSeek, etc.) offer AI-powered research and writing assistance. I have not addressed the capabilities of Google’s new Deep Research tool because it is presently available only to those who pay the $200 per month subscription.
Focus on the Research Process
Brian Schoch teaches business at New Albany-Plain (OH) Local Schools. He spoke to me about a key distinction to keep in mind when we advise students how to use AI tools ethically and effectively for research.
“I think there is currently a large gap in the level of adoption of students, just as there is in teachers,” Shock said.” Some have tiptoed in, others are using it more frequently. With my students, we talk about the difference between using AI for idea generation and utilizing it for an end-result.”
One of my favorite curmudgeons in this thought leadership space is Gary Marcus, who is a professor emeritus of psychology and neural science at New York University. His views on AI’s disruptive effects on teaching and learning are delivered via a newsletter, which is both wickedly entertaining and thought-provoking. Marcus suggests we take the following approach when assigning research tasks to students: “Consider evaluating students on their research process, their ability to identify meaningful questions, and their capacity to recognize when AI tools are and aren’t helpful.”
Why do I think the advice given by these scholars and teachers is valuable for students and teachers? As a professional writer, I do research to inform my writing every day. If these suggestions inform the practice of a professional, surely they can inform the practice of students assigned to do a four-page report on the causes of the Civil War.
The post Redefining Research in a Gen-AI Powered Classroom appeared first on Getting Smart.
Explore the evolving role of research in classrooms with AI, emphasizing critical evaluation, media literacy, and adapting standards for AI-driven education.
The post Redefining Research in a Gen-AI Powered Classroom appeared first on Getting Smart. Personalized Learning, The Future of Tech and Work, Artificial Intelligence Getting Smart
