Max Bohun is founder and CEO of GradeAn Edtech supported by a combinator was there in the “10 Techcrunch startups watching from the W25 Demo Day of YC”.
Education is currently faced with an important challenge: it is not personalized enough to meet the individual needs of students. Overloaded teachers Facing high -class inscriptions, standardized test requirements and administrative tasks.
An investigation in 2022 by Edweek Research Center and the Winston School of Education of Merimack College only found 46% of a teacher's time In school buildings has moved to teaching. The rest is devoted to things such as planning and administrative work, with an average of five hours a week, more than 3 years over a 30 -year career!
As a former educational assistant in Cornell and now as CEO of an Edtech company which helps the note, I knew the challenge of providing personalized comments to large -scale students.
The power of personalized education models
Consider the Oxford tutorial systemWhere students meet tutors once or twice a week in groups of 2-3 to discuss lessons and solve doubts. This human connection can make personal learning, engaging and helping students remain caught.
Why should only the most prestigious institutions benefit from this personalized mentoring? With AI automating lower value tasks, each student can now have access to personalized attention.
The potential of AI in education
I think that the true potential of AI in education is not simply efficiency – it is transformation. By automating administrative tasks, AI can release educators to do what only humans can do effectively: inspire, supervise and connect. AI can make education more fair in several key ways:
1. Immediate and complete feedback
AI systems can provide comments on work for each student, regardless of class size, with a quick turnaround. (Disclosure: My company specializes in this, like others.) This fast feedback loop accelerates learning and provides personalized advice on a large scale.
2. Access to tutoring
Tutors with vocal capacities could provide support for interactive learning, allowing all students – not just those who can pay $ 50 an hour – to obtain tutoring support.
3. Personalized learning course
By analyzing performance data, AI can help identify the unique strengths and weaknesses of each student, allowing a really differentiated instruction adapted to individual styles and learning needs. The Accelerator of the University of Stanford's University is to actively seek how AI can improve education.
4. Commitment of teachers
Most AI tools can allow teachers to concentrate their expertise where it matters most: establishing relationships, developing critical thinking skills and nourishing students' passion for learning. As Victor Lee, the faculty led the AI + education initiative of Stanford notes“I am encouraged to see a certain movement towards the creation of AI tools which improve the lives of teachers – not to replace them, but to give them time to do the work that only teachers are able to do.”
AI key challenges in education
1. Hallucinations and precision
AI systems, in particular language models, can produce “hallucinations” – information with plausible but factually incorrect consonance. These hallucinations occur because IA models Function like Advanced Autoclete tools Designed to predict the following word based on models, not to check the truth.
A study in 2025 examining the research proposals generated by Chatgpt revealed a frequency concerning inaccuracies: of 178 references cited by Chatgpt, 28 did not exist. In a similar study of the medical items generated by Chatgpt, 46% of the references were manufactured and only 7% were authentic and precise.
For the leaders of educational technology who take up this challenge, I found two particularly effective approaches: first of all, implementing human surveillance in loop for critical educational content. Finally, educate students on the limits of AI, teaching them to critically assess the information generated by AI rather than accepting it as authority. I found the most successful implementation position as a starting point for the survey, not the last word.
2. Academic integrity
The emergence of a generative AI has fundamentally disrupted traditional evaluation models. According to a recent survey of students, instructors and administrators, 51% of students would continue to use generative AI tools even if they are prohibited by their institutions or institutions.
Copyleaks research, analyzing data from January 2025 to January 2025, revealed a 76% overvoltage In the equipment generated by AI in the work of students during the year. Interestingly, plagiarism rates experienced a significant decrease of 51% during the same period, which suggests significant challenges in identifying academic dishonesty assisted by AI.
For assignments vulnerable to the generation of AI, educators rotate alternative evaluation methods such as verbal presentations or audio reflections. Many institutions go to pen and paper exams.
According to my experience, the most promising approaches involve an authentic project work which demonstrates the application of knowledge, supplemented by AI tools which facilitate rather than replacing students' thought.
3. Regulatory executives
Educational regulation organizations are focused more and more on several critical concerns concerning the implementation of the AI. The confidentiality of students' data ranks above all – with AI systems to collect and analyze unprecedented quantities of information on the behavior and performance of students. In addition, there are more and more concerns about technological dependence and potential erosion of fundamental cognitive skills when AI tools are implemented without careful railing.
Despite the rapid advancement of AI technologies, school policies are struggling to keep pace. According to a 2025 education week survey, 79% of educators said their districts do not have clear AI policies, creating an uncertain environment for teachers and students.
The California State University system recently announced a initiative Provide its 460,000 students and more equitable access to AI tools, including a partnership with OPENAI to implement Chatgpt Edu. Other states will probably follow, but the implementation must be guided by thoughtful regulatory frameworks which protect the interests of students.
When I take up these regulatory challenges, I advise the leaders of education technology to engage proactively with decision -makers rather than waiting for the regulations to emerge. Develop transparent data governance frameworks that go beyond current requirements and design implementations that center human development rather than technological convenience.
The future of man -centered education
AI in education does not consist in replacing teachers – it is a question of amplifying their impact. With AI, teachers can finally refocus their energy to help students.
Students should expect school impatiently. Teachers should recover their role as mentors rather than administrators. If it is implemented thoughtfully, AI will not make education less human – it could make it more human than ever. But this future forces us to approach these technologies with both optimism and clear recognition of their limits.
Forbes Business Council is the most important growth and networking organization for business owners and managers. Am I qualified?
At Learnopoly, Finn has championed a mission to deliver unbiased, in-depth reviews of online courses that empower learners to make well-informed decisions. With over a decade of experience in financial services, he has honed his expertise in strategic partnerships and business development, cultivating both a sharp analytical perspective and a collaborative spirit. A lifelong learner, Finn’s commitment to creating a trusted guide for online education was ignited by a frustrating encounter with biased course reviews.