What challenges do educational institutions face when implementing AI programs?

Key challenges include lacking adequate infrastructure (people, access, comfort with AI), insufficient digital access and affordability, unclear policies and practices, fear of AI among educators, and a lack of strong data systems for research and development.

How can funding help overcome challenges in AI implementation for educators and students?

Funding can support capacity building for institutions, fund inclusive AI R&D, invest in research and evaluation, help bridge the digital divide by subsidizing access and compute costs, and support the development of ethical guidelines and policies.

What opportunities does AI offer for project-based and constructionist learning?

AI can enable new creative learning experiences, allowing students and educators to create their own tools and express ideas. Platforms that help students find project ideas, create prompts for inquiry, and integrate hands-on resources are key opportunities.

What are the benefits and drawbacks of using AI in personalized learning?

Benefits include AI tutoring that can adapt to individual pace and learning styles. Drawbacks include the need to ensure models are trained on diverse data, the uncertainty about true efficacy compared to decades of edtech efforts, and the risk of AI tutors merely providing answers instead of guiding critical thinking.

How can we promote diversity and inclusivity in AI research and development?

Promote diversity by bringing voices with lived experience to the table through capacity building and cultural shifts in listening. Funding can support inclusive AI R&D, encourage co-design with underserved communities, and advocate for policies that prioritize equity, bias mitigation, and accessibility.

What is the role of AI in future educational assessments?

AI can transform assessment by enabling more formative evaluations and understanding non-cognitive skills. There's a market trend towards AI-driven models, including those for non-cognitive assessments, though defining these competencies and ensuring validity remains a challenge.

How can success be defined and measured for AI interventions in resource-constrained environments?

Success can be defined by improved academic performance, student happiness, and teacher/student well-being. Metrics should consider community-defined success, listening to local needs, and focusing on tangible outcomes like increased agency, economic mobility, and access to quality education, even with limited resources.

What are the concerns about over-quantification and data bias in AI educational tools?

Concerns include the potential for AI to predetermine student paths, privacy issues, and ensuring data used for training models includes diverse populations. There's also a risk of relying too heavily on AI, potentially eroding critical thinking skills, and the challenge of collecting the 'right' data beyond academic metrics.

MIT AI + Education Summit 2024: AI Foundations and Philanthropy

MIT OpenCourseWare

Education5 min read64 min video

Aug 7, 2024|287 views|5

Save to Pod

Key Moments

TL;DR

Funders discuss AI in education, focusing on equity, teacher training, and measurable outcomes.

Key Insights

Philanthropic organizations are investing in AI for education, focusing on capacity building, research, and innovative tools.

Key challenges in AI implementation include lack of infrastructure, insufficient digital access, prohibitive costs, fear of AI, and inadequate data systems.

Measuring the impact of AI funding requires investing in research, elevating evidence standards, and fostering greater learning and transparency within the field.

AI presents opportunities for constructionist and project-based learning by enabling students and educators to create their own AI applications and tools.

Personalized learning through AI offers potential benefits in tutoring and teacher feedback, but drawbacks include the need for more robust research on effectiveness and ensuring equitable access to data.

Funding and policy initiatives can support diversity and inclusivity in AI by co-designing with diverse communities, building their capacity, and establishing accountability measures for companies.

The impact of AI on educational outcomes, particularly non-cognitive skills and engagement, is an emerging area with potential for transformative assessment models, though challenges remain in defining and measuring these competencies.

Success in AI interventions, especially in resource-constrained environments, is shifting towards community-defined metrics and broader social-emotional outcomes beyond traditional academic performance.

PHILANTHROPIC PRIORITIES IN AI AND EDUCATION

Panelists from major philanthropic organizations detailed their current funding priorities at the intersection of AI and education. These include integrating computational thinking into curricula, enhancing reading and writing skills through personalized learning, and leveraging generative AI to accelerate these strategies. Investments are directed towards capacity building for institutions and governments to understand and implement AI, as well as research and validation of new AI tools and models. Specific focuses include AI tutoring, teacher coaching tools, and applications for differentiated student learning, with a significant emphasis on generating evidence of effectiveness.

CHALLENGES IN IMPLEMENTING AI LITERACY AND PROGRAMS

Implementing AI literacy and other AI-driven programs in educational institutions faces several significant hurdles. These include a lack of essential infrastructure, insufficient digital access for many students, and prohibitive costs associated with AI tools and compute power. Furthermore, there's a prevalent fear of AI among educators and a lack of sophisticated data systems for research and development. Insufficient policies and practices, alongside a failure to integrate AI into existing systems rather than creating new verticals for it, also pose substantial challenges. Overcoming these obstacles requires significant investment in infrastructure, inclusive R&D, and robust measurement and evaluation frameworks.

MEASURING THE IMPACT OF AI FUNDING ON EDUCATIONAL OUTCOMES

Measuring the impact of AI funding on educational outcomes is crucial for accountability and evidence-based practice. This involves investing in research to understand the efficacy of new AI tools, particularly for teacher coaching and AI tutoring. Elevating evidence standards in the edtech field, akin to ESSA standards, is essential to ensure quality and effectiveness. Creating more spaces for shared learning, fostering transparency, and lifting up insights across the field are also vital. This approach aims to normalize and standardize the measurement of AI interventions, considering not only student and teacher outcomes but also adherence to ethical guidelines and data privacy standards.

ENABLING CONSTRUCTIONIST AND PROJECT-BASED LEARNING WITH AI

AI holds significant potential to foster more constructivist and project-based learning experiences. Platforms that empower both students and educators to create their own AI applications and tools are key. This approach shifts from a deficit model of assessment towards equipping learners to express ideas, work with communities, and develop their own solutions. The focus is on enabling children to create projects, find ideas for projects, and formulate effective prompts for inquiry. While gamification can make learning fun, the goal is to leverage AI to support deeper, hands-on creation and critical thinking, rather than solely for assessment or efficiency.

THE ROLE OF AI IN PERSONALIZED LEARNING: BENEFITS AND DRAWBACKS

AI plays a significant role in personalized learning, particularly through AI tutoring and AI-driven feedback for teachers. Potential benefits include tailoring learning pace and methods to individual student needs, and providing teachers with valuable insights into their practice. However, significant drawbacks exist, including the early stage of research on the effectiveness of AI tutors, challenges in incorporating data from underserved students into AI models, and the risk of AI tutors simply providing answers rather than guiding students through questioning. Philanthropy's role is key in supporting research to identify what works, for whom, and under what conditions, and in generating diverse datasets for model training.

PROMOTING DIVERSITY AND INCLUSIVITY WITH AI FUNDING AND POLICY

Supporting greater diversity and inclusivity in AI research and development requires a multifaceted approach. Funding and policy initiatives should focus on bringing voices with lived experience to the table, not just numerically but by building their capacity and fostering a culture of listening. Capacity building extends to educators and communities, empowering them to develop their own data sets and platforms. Accountability measures for AI companies are needed to ensure equity, mitigate bias, and ensure accessibility. Co-creation with diverse communities, collaboration between researchers, school districts, and industry, and a focus on understanding how AI impacts diverse student populations are critical for equitable AI development.

ASSESSING NON-COGNITIVE SKILLS AND STUDENT ENGAGEMENT WITH AI

The landscape of AI-driven assessment for non-cognitive skills and student engagement is emerging, with potential for formative assessment models. While platforms are experimenting with integrating AI into assessments to understand student perceptions and engagement, challenges remain in defining and measuring these competencies universally. There's a risk of over-quantification and a focus on efficiency plays that entrench current paradigms rather thantransforming them. Moving forward, success will involve developing common taxonomies, measuring skills beyond academics, and potentially leveraging AI to process data that reflects student well-being and engagement throughout the learning day.

DEFINING SUCCESS IN RESOURCE-CONSTRAINED ENVIRONMENTS

Defining success for AI interventions in resource-constrained environments requires a shift towards community-defined metrics and broader outcomes. While quantifiable academic achievements are important, success is increasingly viewed through a lens of social-emotional impact and the joy of learning. This necessitates listening to communities, understanding their needs, and facilitating conversations to address local barriers. Success metrics can range from high-level goals like increased opportunities and lifetime earnings for students to portfolio-level indicators of engaging math products and grant-specific benchmarks for AI tool performance, ultimately aiming to move the needle on meaningful student outcomes.

Mentioned in This Episode

●Software & Apps

●Companies

●Organizations

●Concepts

●People Referenced

Navigating AI in Education: Dos and Don'ts for Institutions and Funders

Practical takeaways from this episode

Do This

Invest in and build robust digital infrastructure (people, tools, access).

Support inclusive AI research and development that centers community voices.

Develop clear policies for AI implementation and ethical use.

Prioritize teacher professional development and comfort with AI tools.

Focus on AI tools that enhance, rather than replace, human interaction and critical thinking.

Invest in research to understand AI efficacy and evidence standards (e.g., Essa standards).

Promote co-design processes involving researchers, educators, students, and product developers.

Consider AI's potential for formative assessment and understanding non-cognitive skills.

Ensure transparency and choice regarding data collection and usage.

Encourage the development of community-created AI datasets and platforms.

Avoid This

Don't solely rely on outside expertise without building internal capacity.

Don't overlook the digital divide; ensure equitable access to technology and AI tools.

Don't implement AI without considering prohibitive costs for families.

Don't ban AI usage outright; develop thoughtful policies instead.

Don't ignore the fear educators and students may have towards AI.

Don't neglect data privacy and ethical guidelines when developing or deploying AI.

Avoid AI tools that solely focus on efficiency or entrenching current paradigms without transformative potential.

Don't focus solely on academic outcomes; consider broader measures like engagement, motivation, and well-being.

Avoid AI models that predict a fixed path for students without considering their potential for growth.

Do not solely collect data for academic metrics; gather diverse data including perceptions and non-cognitive skills.

Common Questions

Funding priorities include building AI infrastructure, supporting inclusive AI R&D, investing in research to understand AI efficacy, developing AI tools for personalized learning and teacher support, and promoting AI literacy programs.

Topics

Funding Strategies Equity In AI Future Of Learning

Mentioned in this video

organizationOverdeck Family Foundation

Invests in providing opportunities for all children to unlock their potential, with a focus on early childhood, out-of-school STEM, teacher professional learning, and student-centered learning. They emphasize evidence-based research and have invested significantly in AI tools and research for education.

softwareQuill

Mentioned as an example of a writing application that uses AI for student-facing interventions and to improve outcomes.

personGeorge Siemens

Credited for the term 'human-AI entanglement,' a concept discussed in relation to supercharging the capabilities of teachers and systems with AI.

organizationRobin Hood Foundation

Funds initiatives in New York City focused on economic mobility through technology and learning, aiming to reach a million students and 1,800 schools. Their Learning and Technology Fund focuses on computational thinking and improving reading/writing skills.

organizationLego Foundation

A corporate foundation focused on children's development and learning through play, aiming to equip children with creativity and critical thinking skills for lifelong learning. They have shifted to a more community-oriented approach and focus on system-level challenges.

organizationHong Kong Jockey Club Charities Trust

A non-profit organization that donates significantly to various initiatives including education, poverty alleviation, and talent development. Their 'Cool Think' project, developed with MIT, focuses on equipping the next generation with 21st-century skills, including AI.

organizationFederation of American Scientists

An independent organization through which the Lego Foundation funds US government research fellows to build capacity in understanding AI and technology.

organizationPoliU Department of Computing

Daniel Li is a research team member here, developing AI and big data models to predict student learning and behavioral outcomes.

softwarePlaylab.ai

An application developed by an MIT alum, a member of the Scratch team, that enables educators to create their own AI applications for educational purposes.

organizationBlue Ridge Labs

A division of Robin Hood that has a design insight group with 2,500 people from New York City with lived experience, collaborating with tech innovators to ensure user design incorporates diverse perspectives.

organizationThe Gates Foundation

Invests in education, with a specific focus on US K-12 math education due to its predictive power for high school graduation and future success. They support innovative math curricula and teacher professional development, exploring AI's potential for tutoring and teacher support.

organizationEmpire AI

A recently launched collaboration in New York involving seven higher education institutions and government, aiming to build AI hubs and incorporate diverse voices into AI development.

toolPisa

toolKhan Academy