The ongoing debate about artificial intelligence (AI) in education highlights its potential to either bridge or exacerbate existing inequalities. Following the launch of ChatGPT in late 2022, New York City’s largest public school system quickly moved to ban the chatbot, citing concerns over its potential negative impacts on student learning and issues related to content security and accuracy. In contrast, the private Franklin School in Jersey City, New Jersey, embraced AI as a vital aspect of its curriculum, aiming to enhance educational experiences rather than replace teachers.
Will Campbell, the head of Franklin School, explained that the approach focuses on enriching student learning while also creating efficiencies for teachers. The school, which opened in 2022, integrated AI tools to offload routine administrative tasks, allowing educators to dedicate more time to instruction. Early AI initiatives included custom chatbots trained on approved course materials, serving as tutoring aids that fostered deeper student engagement.
At the university level, Ethan Mollick, a professor at the Wharton School of the University of Pennsylvania, has championed AI use in education since January 2023, incorporating explicit guidelines into his syllabus. A vocal advocate for AI, Mollick collaborates with AI companies like OpenAI to create educational materials. He emphasizes that AI holds substantial promise in addressing significant educational challenges, stating, “We have some early evidence that it’s an incredibly powerful teaching tool.”
Diverging opinions on AI’s role in education persist. Some educators argue that AI will never replace the benefits provided by human teachers, while others contend that AI-powered tutoring offers a valuable alternative, particularly in addressing access disparities. The consensus is clear: students benefit from tutoring, yet equitable access remains a pressing issue.
The “tutoring gap” has long been recognized, with decades of research indicating that individualized tutoring significantly enhances student learning outcomes. However, personalized, at-home instruction remains unattainable for many families due to staffing shortages and high costs. AI tools now promise a tutor-like experience that students can access from home.
In her analysis, Jennifer Steele, a professor at the School of Education at American University, highlighted the challenges families face in affording one-on-one tutoring, especially given the high salaries teachers command. A peer-reviewed study from the University of Wisconsin-Madison in 2024 revealed a strong correlation between socioeconomic status and access to tutoring. The University of Southern California’s Understanding America Study found that only about 15% of students received any form of tutoring, with fewer than 2% accessing even moderately high-quality instruction. Alarmingly, among students earning grades of C or lower—those most likely to benefit from additional support—fewer than 4% received high-quality tutoring.
AI has the potential to address this access gap. With advances in natural language processing, AI tools can engage in conversational interactions, explain complex concepts, and offer tailored responses that mimic human tutors. Students can upload images of their work or share live video questions, extending the support they receive. Michael Hilton, a professor at Carnegie Mellon University, noted a significant drop in office hour attendance as students increasingly turn to AI for assistance. This shift allows educators to focus on higher-level concepts rather than basic queries, creating a more enriching learning environment.
The cost of traditional tutoring remains a significant barrier, with a report from the Brookings Institution estimating that effective tutoring programs can require between $1,000 and $3,000 per student annually. The disparity in access to private tutoring creates further inequalities in educational achievement.
Aly Murray, an advocate for low-income students and founder of Upchieve, an edtech nonprofit, shares firsthand experience of these disparities. Upchieve provides free, 24/7 online tutoring and college counseling to low-income middle and high school students nationwide. While the platform incorporates AI for specific tasks, Murray emphasizes that human tutors remain irreplaceable in the tutoring process.
Despite AI’s promising applications, recent research by Upchieve indicates limitations. A study involving the UPbot, an AI tutoring chatbot trained on extensive human tutoring transcripts, revealed low engagement levels among students. Only one in five students utilized the AI tool, while 92% of tutoring sessions involved human tutors. Students expressed a strong preference for human interaction, valuing the relational aspects of learning that AI cannot replicate.
The challenges of AI in education extend to its effectiveness. Upchieve’s study found no statistically significant differences in learning outcomes or student confidence between sessions led by human tutors and those conducted through AI. This raises concerns about AI’s capacity to independently close educational gaps.
Murray warns that while AI can support low-income students, it may also widen existing inequalities if affluent students leverage these tools more effectively. This phenomenon recalls historical patterns of technological adoption, where disparities often accompany new advancements. For example, a study using data from the City Health Dashboard revealed that high-income neighborhoods had a broadband access rate of 87.2%, compared to just 58.8% in low-income neighborhoods.
The prevailing sentiment among education experts is that while human tutoring remains the ideal, the critical question is whether AI tutoring provides better support than no assistance at all. Campbell argues that deploying AI in communities lacking resources could be a significant step forward for education.
Companies in the edtech sector are grappling with these complexities. McGraw Hill, a long-established educational publisher, has evolved to incorporate AI in a manner that prioritizes human learning experiences. Dr. Dylan Arena, McGraw Hill’s chief data science and AI officer, highlights the importance of maintaining a human presence in the educational process, allowing AI to enhance rather than replace traditional teaching methods.
Furthermore, platforms like Khan Academy have integrated AI into their offerings with tools like Khanmigo, designed to support learning rather than simply provide answers. Dr. Kristen DiCerbo, Chief Learning Officer at Khan Academy, emphasizes the platform’s mission to ensure equitable access to educational resources, particularly for students in diverse backgrounds.
AI’s role in education continues to evolve, with numerous tools emerging to enhance learning experiences. These include features from various AI chatbots that adopt the Socratic method, encouraging students to engage deeply with their studies rather than simply seeking answers.
As the education landscape transforms, the challenge remains: how can we harness AI to close the educational gap without inadvertently reinforcing existing disparities? The pursuit of equitable solutions in this space will be crucial as stakeholders navigate this complex terrain.
