A new study by the Brookings Institution's Center for Universal Education indicates that the risks associated with using generative artificial intelligence (AI) in K-12 education currently outweigh its benefits. The report, based on focus groups, interviews with students, parents, educators, and tech experts in 50 countries, and a review of hundreds of research articles, describes AI's potential damages as "daunting" but "fixable." The authors characterize their review as a "premortem" to examine AI's classroom potential without the benefit of long-term data or hindsight. Some key findings include:
Benefits of AI in Education
- Enhanced Learning Support: Teachers found AI useful for language acquisition, especially for second-language learners, by adjusting passage complexity and offering privacy. It can also improve students' writing by sparking creativity, aiding organization, and refining grammar during drafting and revision, provided it supplements, not replaces, teacher efforts.
- Teacher Efficiency: AI can automate various administrative tasks for teachers, such as generating parent emails, translating materials, and creating worksheets, rubrics, quizzes, and lesson plans. One U.S. study cited found that teachers using AI save nearly six hours per week.
- Equity and Accessibility: AI has the potential to reach excluded children, exemplified by a program in Afghanistan digitizing curricula for girls via WhatsApp. It can also make classrooms more accessible for students with learning disabilities like dyslexia.
Risks of AI in Education
- Cognitive Development Threats: The report highlights concerns that AI can negatively impact children's cognitive growth. Increased dependence on AI for answers may lead to cognitive off-loading, reducing students' ability to think critically, parse truth from fiction, or engage deeply with material. Evidence suggests declines in content knowledge, critical thinking, and creativity among students who use generative AI.
- Social and Emotional Development Threats: Overuse of AI, particularly chatbots, may undermine students' emotional well-being, their ability to form relationships, recover from setbacks, and maintain mental health. Chatbots, designed to reinforce user beliefs, can create an "echo chamber" that makes students uncomfortable with real-world disagreements. A survey found nearly 1 in 5 high schoolers reported a romantic relationship with AI, and 42% used AI for companionship.
- Exacerbated Inequity: While AI can promote equity, it can also worsen existing disparities. Free AI tools are often less reliable and accurate than more advanced, paid models. This means schools in wealthier communities may access superior AI, creating a situation where accurate information requires greater financial resources, disadvantaging underfunded districts.
Recommendations
The Brookings report offers several recommendations for parents, teachers, policymakers, and tech companies:
- Shift Educational Focus: Schools should prioritize fostering curiosity and a desire to learn over "transactional task completion" to reduce student reliance on AI.
- Rethink AI Design: AI tools for children and teens should be less sycophantic and more "antagonistic," challenging users to reflect and evaluate.
- Promote Collaboration: Tech companies should engage with educators in "co-design hubs" to develop and test new AI applications, mirroring initiatives like the government-backed hub in the Netherlands.
- Implement AI Literacy: Comprehensive, national AI literacy guidelines for both teachers and students are crucial, similar to frameworks in countries like China and Estonia.
- Address Funding Disparities: Efforts must ensure underfunded districts are not left behind in AI adoption to prevent further educational inequity.
- Enact Government Regulation: Governments are responsible for regulating AI use in schools to safeguard students' cognitive and emotional health, as well as their privacy.
The authors emphasize the urgency of addressing AI's current risks and implementing remedies now.