OpenAI Warns Students Against Using ChatGPT as an 'Answer Machine' and Pushes for Productive Struggle in AI Education

OpenAI’s head of education, Leah Belsky, is urging schools to rethink how students use AI tools like ChatGPT, warning against treating them as simple "answer machines." Speaking on OpenAI’s podcast, Belsky emphasized that AI should not replace the learning process but enhance it when used thoughtfully. She stressed that students must learn to use AI responsibly and creatively, starting early in their education. Belsky, who has led OpenAI’s education initiatives since 2024, argued that AI is fundamentally a tool—much like a calculator—and its value depends on how it’s applied. If students rely on it to get answers without engaging in critical thinking, they miss out on real learning. “What matters most is how that tool is used,” she said. “If students use AI as an answer machine, they are not going to learn.” Instead, OpenAI is working to embed “productive struggle” into AI-powered learning. This concept refers to the idea that learning happens through effort, challenge, and problem-solving—not just by receiving correct answers. To support this, OpenAI recently launched “Study Mode” in ChatGPT, which offers guiding questions tailored to a student’s skill level and learning goals. The feature is designed to encourage deeper understanding by prompting students to think through problems rather than passively accepting responses. Belsky also highlighted the importance of coding literacy, calling it a foundational skill for the future. She pointed to “vibe coding”—using natural language prompts to generate code—as a powerful new method, especially as AI tools make programming more accessible. However, she noted that users still need some technical knowledge to review and correct AI-generated code, which helps maintain learning and accountability. Other companies are also addressing this challenge. Kira Learning, co-founded by Google Brain’s Andrew Ng, focuses on helping non-technical teachers deliver computer science education. Its AI agents are designed to introduce deliberate friction at key moments, encouraging students to wrestle with concepts before turning to AI for help. The broader issue remains: many schools still treat AI use as cheating and focus on preventing it, rather than teaching how to use it effectively. Tyler Cowen, an economics professor at George Mason University, criticized this mindset, saying the education system is overly focused on grades rather than real learning. “The whole system is set up to incentivize getting good grades. And that's exactly the skill that will be obsolete,” he said. As universities and schools struggle to adapt, tech companies like OpenAI and Kira are stepping in to shape the future of AI-assisted education—making sure that learning remains meaningful, even in an age of intelligent machines.