Research

4 Keys to Exploring AI in Education

• By Kate Lucariello
• 03/25/24

As part of its recent "Reimagine Education" 2024 conference, Microsoft Education released its "AI in Education" special report containing research insights, resources, and recommendations for tapping into the advantages — and caveats — of using AI.

While the report acknowledges AI's advantages and utility for students, teachers, administrators, and IT professionals, it also emphasizes "we're just beginning to understand the future of AI in education" and the need to address current and future challenges of AI responsibility, trust, and transparency.

The report is divided into four key takeaways:

Start AI conversations today: According to a Microsoft survey of 1,800 K-12 and higher education students, educators, and academic and IT leaders, while the majority of respondents have used AI in school at least once or twice to "enhance productivity, personalize learning, and improve efficiency," a significant portion (24% of leaders, 42% of educators, and 52% of students) cite plagiarism as the top concern about using it. Recommendations from the report include:

• Talk to students about how they can use AI without compromising academic integrity;
• Create clear guidelines and policies; and
• Include AI literacy in class instruction and staff training.

Learn how AI can help: Institutions are seeing over a threefold return on their investment within two years of AI adoption, reported a March 2024 IDC study, but there is a need for a coordinated adoption strategy aligned with institutional objectives. Recommendations in this area include:

• Appoint an oversight committee to define strategy;
• Create a "centralized, cross-functional team" to align AI adoption with existing policies; and
• "Prioritize high-value, low complexity AI use cases" to pace adoption and use of AI and plan for solutions as needs arise.

Explore new ways to learn: In a randomized experiment on large language models and education, Microsoft Research and Harsh Kumar from the University of Toronto found that AI tools that offer tutoring and encourage engagement rather than simply giving answers work best, especially when students attempt to solve problems on their own before consulting the AI. Recommendations in this area include:

• Foster a "growth mindset" of "learning, iteration, and curiosity;"
• Talk to others using AI in education to get ideas and support; and
• Focus on goals in preparing AI tools and how AI might help achieve them.

Prepare for the future: Microsoft's 2023 Work Trend Index survey found that 82% of business leaders believe employees will need new AI skills for their careers, with the top skill being "analytical judgment" about AI outputs, followed by people skills. Recommendations in this area include:

• Educate students in "metacognitive skills," i.e., the ability to understand why they agree or disagree with an AI's output;
• Use AI tools to "encourage evaluation rather than provide answers" and to know when to delve deeper; and
• Develop "human-centered skills," as businesses value them in addition to tech skills.

The full AI in Education report can be downloaded here.