OpenAI introduced a major update to ChatGPT on Tuesday with the launch of Study Mode, a new feature designed to help students learn, rather than just giving them answers.
This is the latest sign that education is becoming a new battlefield in the AI war.
ChatGPT has been seen as a cheating tool in the past because it helped students easily knock out essays and other homework without much work. Study Mode tackles this by purposely injecting friction into the process — friction that’s needed for humans to actually learn.
The move signals a strategic push by OpenAI to position its flagship AI product not just as a productivity tool, but as an educational companion that could help the company attract and retain more younger users.
“When ChatGPT is prompted to teach or tutor, it can significantly improve academic performance. But when it’s just used as an answer machine, it can hinder learning,” Leah Belsky, VP of education at OpenAI, said.
Younger mindshare
OpenAI sees Study Mode as part of a broader educational vision.
“For educators, AI can free up time for the human work of teaching. For institutions, AI will become core infrastructure, like the internet, reshaping how we teach, research, and run operations,” Belsky added. “For students, AI holds the most powerful potential of all, the ability to serve as a personal tutor that never gets tired of their questions.”
Winning the mindshare of a younger audience is key for Big Tech companies because, as these people grow up and enter the workforce, they often take the tools they learned to use in school and spread them across companies and the broader society.
Google mastered this technique years ago when it got young students hooked on free workplace apps, including Google Docs, Sheets, and Slides. Now, these users are older and have brought this software into more official work situations, where companies often pay for access through Google’s Workspace service.
Google vs. OpenAI, with Khan in the mix
A replay of this strategy is beginning to play out in the generative AI arena, where OpenAI, Google, and others are racing to redefine digital education through more personalized AI learning experiences.
Last year, Khan Academy, a pioneering online education provider, launched Khanmingo, an AI-powered assistant for teachers and students that uses OpenAI technology.
Earlier this year, Google announced Gemini for Education, which offers tools such as bespoke quizzes, AI-generated lesson plans and simulations, along with study video overviews. On Tuesday, the company unveiled new education features that are woven into its new AI Mode in Google Search.
Last week, OpenAI unveiled a tighter integration with Canvas, the leading learning app used by thousands of schools and colleges in the US. This week, the startup launched Study Mode. It’s available to all logged-in users of ChatGPT Free, Plus, Pro, and Team plans starting on Tuesday. It will roll out to ChatGPT Edu in coming weeks.
From answers to learning
Unlike traditional interactions with ChatGPT, which often deliver quick, complete answers, Study Mode actively avoids this, and instead guides students along a learning journey by asking them to clarify questions, while gauging their skill levels and tailoring explanations, quizzes and other responses accordingly.
In Study Mode, if a student tells the chatbot to just give the answer, the technology will respond by reminding them they’re here to learn, and then it will try to entice them back into the learning flow with another question or suggestion.
The technology incorporates Socratic questioning techniques, personalized feedback, and a knowledge check, with a goal of encouraging learners to reflect, struggle, and engage deeply — trying to mimic the role of a patient human tutor.
The feature was built with input from teachers, learning scientists, and pedagogy experts from about 40 institutions, according to OpenAI. The underlying system instructions were designed to foster curiosity and help students understand the material rather than memorize answers.
Study Mode’s current implementation uses custom system instructions rather than deep AI model training. This helped OpenAI release the first version of Study Mode quicker, and will let the company iterate on the product based on more student feedback. The startup said it plans to incorporate these behaviors into its core models over time, signaling an even deeper integration of pedagogy into AI design.
Read the full article here
