What are the main challenges and opportunities for teachers when using generative AI?

Challenges include concerns about cheating and the need to rethink learning outcomes. Opportunities lie in using AI for curriculum development, personalized learning, and preparing students for a workplace where AI tools are common.

How can generative AI be used to improve the coherence and accuracy of educational materials?

Dr. England explained how LLMs can automate lecture updates, ensuring consistency across series. He also detailed efforts to develop 'automated interpretable reasoning engines' (AIRE) to reduce hallucination rates and enhance truthfulness in AI outputs.

What is 'Backward Design' and how does it apply to integrating AI in education?

Backward Design involves starting with learning goals, then revising assessments and assignments, and finally redesigning teaching activities. Janet Ranken suggests using this framework to determine if AI necessitates a shift in educational goals or assignment structures.

What is the role of educators in teaching students about generative AI?

Educators have a responsibility to help students understand AI literacy in a contextualized manner, focusing on responsible use. Eric Kloer compares this to media literacy, emphasizing that AI literacy needs to be integrated across different subjects.

How will generative AI impact the role of Teaching Assistants (TAs)?

The panel suggested TAs may become even more valuable by focusing on contextual understanding, evaluating AI output, and providing the human/social aspects of teaching. AI could automate some monotonous tasks, freeing TAs for more impactful interactions.

What advice is given to educators who are hesitant to embrace constructionist approaches and AI?

Panelists encourage educators to be honest about what they know regarding student learning, leverage the science of learning and cognitive research, and accept that change is often incremental. They stress the importance of active learning where students construct meaning.

Key Moments

GenAI + Education Reinvent Teaching

MIT OpenCourseWare

Education3 min read41 min video

Dec 11, 2023|1,730 views|18

Save to Pod

Key Moments

TL;DR

Generative AI is being integrated into education to reinvent teaching, focusing on design, assessment, and teacher professional development.

Key Insights

Generative AI necessitates a fundamental rethinking of learning outcomes and assessment methods.

Teachers require new skills and professional development to effectively integrate AI into their teaching practices.

AI can be used constructively, enabling students to create and innovate, rather than solely focusing on avoiding cheating.

The workplace is already utilizing AI, making it imperative for students to learn these tools to remain competitive.

AI tools can assist educators in generating examples, quizzes, and summarizing student work, freeing up time for higher-level cognitive tasks.

Encouraging a constructionist approach to teaching and learning is key to embracing AI's potential while mitigating risks.

THE ROLE OF THE TEACHER IN THE AI ERA

Educators are at the forefront of integrating generative AI into classrooms, shifting from traditional teaching methods to a focus on responsible AI use. Initiatives like 'Responsible AI for Social Empowerment and Education' (RAISE) aim to empower students to create and share with AI. This requires preparing teachers through curriculum development and professional training, emphasizing that teachers need to understand AI for both their professional practice and for guiding students. The focus is moving towards reimagining educational goals in light of AI's capabilities, rather than solely combating issues like cheating.

RETHINKING CURRICULUM AND ASSESSMENT

The advent of generative AI compels a re-evaluation of what students need to learn and how their learning is assessed. Instead of fighting against AI-assisted plagiarism, educators are encouraged to rethink assignments and learning outcomes. If AI can easily complete a task, it signals a need for students to engage in higher-order cognitive processes such as synthesizing, evaluating, and creating. The key is to adapt assessments to measure these advanced skills, ensuring students are prepared for a workplace that increasingly relies on AI tools.

INTEGRATING AI IN COMMUNICATION AND DATA COURSES

In subjects like communication and data analysis, generative AI presents opportunities to enhance learning. Instructors are experimenting with requiring students to use AI tools for assignments, like drafting cover letters or generating synthetic data. This allows students to understand AI's capabilities and limitations, while also focusing on the critical aspects of evaluating AI output and communicating effectively in a professional context. Early observations suggest that strong writers often produce better AI-generated content, highlighting the interplay between AI proficiency and foundational skills.

DEVELOPING TRUTHFUL AND INTERPRETABLE AI TOOLS

The development of AI tools for education raises questions about their reliability, particularly the issue of 'hallucinations' or incorrect outputs. In technical fields like physics and mathematics, where truth is often binary, there's a drive to create AI systems that not only generate answers but also provide verifiable reasoning. Projects are focusing on building 'automated interpretable reasoning engines' that can generate code, check results, and ensure a near-zero hallucination rate by mapping problems to known truthful functions, moving beyond 'unintelligible artificial intelligence'.

LEVERAGING AI FOR TEACHER PROFESSIONAL DEVELOPMENT

Generative AI can significantly support teachers' professional development and classroom practice. AI tools can help educators create diverse examples for learners, generate quick quiz questions, and summarize student feedback, thus freeing up valuable time. This allows teachers to focus more on nuanced instruction, strategic thinking, and engaging students in active, constructivist learning experiences. By embracing AI assistively, educators can elevate their teaching, moving towards higher levels of Bloom's Taxonomy and fostering deeper student engagement.

THE FUTURE OF TEACHING ASSISTANTS AND EDUCATOR RESISTANCE

The integration of AI into education prompts a discussion about the evolving roles of teaching assistants (TAs). Far from diminishing their value, AI may empower TAs to focus on more impactful tasks, such as providing contextual understanding, critical evaluation, and personalized student interaction, moving away from repetitive duties. Overcoming resistance to pedagogical shifts, like embracing constructivism, requires leveraging the science of learning, being honest about student learning processes, and updating institutional assessments to reflect the new landscape where AI is an integral part of learning and professional practice.

Mentioned in This Episode

●Products

●Software & Apps

●Companies

●Organizations

●Concepts

●People Referenced

Common Questions

Panelists discussed piloting assignments using AI, redesigning curriculum with AI in mind, and using AI as a tool to enhance communication and reasoning skills. They emphasize hands-on exploration and adapting teaching goals based on AI capabilities.

Topics

Teaching Practice Backward Design AI In STEM Learning Outcomes

Mentioned in this video

People

Eric Kloer

Professor of Education at MIT and Director of the SHetter Teacher Education Program, discusses the RAISE initiative and preparing teachers for AI in education.

Melissa Webster

Lecturer in Managerial Communication at MIT Sloan School of Management, shares her experimental approach to integrating generative AI into her communication courses.

Janet Ranken

Director of MIT's Teaching and Learning Lab, provides insights on adapting teaching practices and embracing generative AI through backward design.

Dirk England

Associate Professor of Electrical Engineering and Computer Science at MIT, discusses developing AI and quantum hardware and using LLMs for educational tools.

Marin Suich

Mentioned as a colleague working with Dr. England on developing an automated interpretable reasoning engine ('AIRE').

Wiggins and McTighe

Authors credited with developing the Backward Design methodology for curriculum development, referenced by Janet Ranken.

Jeremy Wordheimer

Mentioned as collaborating with Dr. England and other MIT community members on developing an AI reasoning engine.

Ethan Mik

An individual whose work on using generative AI for the science of learning is mentioned by Melissa Webster and Janet Ranken.

Organizations

Teaching and Learning Lab

Janet Ranken's affiliation at MIT, which works with faculty on teaching practices and integrating new educational tools and methodologies.

Responsible AI for Social Empowerment and Education (RAISE)

An initiative focusing on empowering children to create and innovate with AI technologies, emphasizing classrooms as a key gateway for equitable access and understanding.

Sloan School of Management

Melissa Webster's affiliation where she teaches communication courses, discussing how the school's curriculum is being adapted to incorporate generative AI.

Companies

Light Matter

A company founded by Dr. England's students, aimed at creating accelerated tools for training large language models and other neural networks.

Concepts

Backward Design

A methodology for designing classes by defining goals, revising assessments, and then redesigning teaching and learning activities, proposed as a framework for integrating generative AI.

Products

AI accelerators

Hardware developed to power future AI tools, with Dr. England's research group focusing on this area, including for quantum computers and large language models.

Software & Apps

MIT Scratch

A platform used to help children build and share creative projects with AI, mentioned as a tool for empowering students in educational settings.

Day of AI