Inclusive AI Literacy & Learning
We’re RAICA a team of MIT’s Artificial Intelligence experts, researchers, curriculum developers, and project-based learning aficionados on a mission to set the standard for how middle schoolers learn about AI. We aim to grow students’ skills as informed consumers and ethical producers of AI tools and technology and to foster their confidence to change the world through computational action. Our curriculum is project-centered, which means students will produce authentic learning artifacts by the end of each module and their learning will be driven by exploring and doing. They will progress through their projects using design thinking steps, and along the way they will deepen their understanding of AI literacy, computational action, and ethical thinking. Please click below if you’re an educator interested in piloting at your school.
Children today are more than digital natives who grow up immersed in technology, they are increasingly becoming AI natives, as well. They interact with AI-powered products and services pervasively in social media, digital assistants, search engines and smart devices often without being fully aware of AI’s presence, how it works and shapes their lives and their futures. Inclusive AI Literacy and Learning — a groundbreaking new research program at MIT — aims to prepare students with an important foundation of AI awareness, skills, critical thinking and attitudes to become responsible users and makers of AI-enabled technologies. Through our work, we hope children will also learn to become ethical and empathetic designers and implementers of AI-enabled solutions and will learn how to respond effectively to its many opportunities and risks.
A key mission of this Program is inclusion, which is defined on three dimensions:
Accessibility: the learning experiences should offer multiple modes of engagement and expression to be accessible to students of diverse cognitive and physical abilities;
Equity: the curriculum seeks to serve students from many cultural, ethnic, and socioeconomic backgrounds whose needs may otherwise be marginalized, ignored or excluded; and
Adaptability: curriculum and activity design should include modular materials that may be adapted to a broad range of learning environments, resources availability, teacher-to-student ratios, and levels of family involvement.
In other words, we aim through our work to encourage children of all backgrounds and abilities and empower their teachers, parents, caregivers and peers with appropriate tools and methodologies to create and sustain a nourishing and empowering learning experience.
Our research Program currently offers two main areas of study for their impact on children’s learning:
How does this inclusive AI curriculum prepares students to be responsible, AI-aware community members?
We are currently focusing on middle school and we call this curriculum Responsible AI for Computational Action (RAICA). Initially, we will develop this curriculum along with supporting hands-on learning projects, tools, technologies, and teacher professional development for middle school. Over time, we plan to extend this to primary, middle and secondary schools, and will make our work available for broad use worldwide.
How might child-friendly intelligent learning companion technologies support learning by personalizing and adapting to specific learning needs?
We focus on the application of AI in early childhood for communication and literacy development using emotionally-engaging social robots.
Inclusive AI Literacy for Middle School
To develop and implement RAICA and to support our inclusivity mission, we are eager to be collaborating with Dubai Heights Academy, an inclusive international school in the UAE along with a variety of middle schools in the U.S. that all specialize in supporting children with a wide range of learning needs. Emphasizing support for a diversity of learners with a range of interests and learning needs (cognitive, social, etc.), we expect to prepare students to create meaningful projects with AI technologies through ethical design and implementation practices. Employment opportunities for those with competence in innovative technologies and creative and critical thinking abilities are very promising across a variety of industries. The future holds a wide range of opportunities for those with skills in computational thinking, artificial intelligence, human-centered computing, and design thinking.
Our pedagogy takes a constructionist approach where students are active learners who work with peers on creative projects that have personal meaning and benefit their community. Students “learn-by-making”, supported and scaffolded by teachers as knowledgeable guides and mentors, to acquire technical skills and concepts in AI and computational thinking while also learning important teamwork and 21st-century skills such as collaboration, communication, creative problem-solving, critical thinking, and design thinking. To support this, we are developing and providing students with intuitive block-based coding tools and other technologies so they are empowered to make computational projects with exciting AI capabilities. As this curriculum is under development, we will be iterating and revising our materials, incorporating feedback from students and teachers, while being mindful of how to adapt them to support students with diverse learning needs. We will also welcome student and teacher participation as co-designers to make sure these activities are relevant and engaging for everyone. All this hard work will result in the first curriculum of its kind that will ultimately be available to other teachers, students, and schools all over the world to promote a more inclusive and equitable future with AI.
Personalized Learning Companions for Early Childhood Education
In this Program we will also be exploring and investigating how AI-enabled technologies strengthen young children’s socioemotional and language learning and development. Specifically, we will be exploring how socially interactive robots support personalized learning in early childhood (focusing on children 4 to 6 years old) for communication and early literacy skills. We are working with Dubai Heights Academy, as well as with schools in the U.S. including underserved schools.
Our prior research has shown that children with diverse learning needs respond well to playful, pet-like, learning companion robots. Our robots are designed to play educational games with children to enhance what children traditionally gain in the classroom with teachers. At MIT we’ve been able to show significant learning improvement in vocabulary and early literacy skills, for instance, by having the robot automatically personalize its interactions with young children. We are particularly interested in exploring how this personalization supports child learners with diverse learning needs. We will also explore new learning experiences, such as augmented reality, which may also augment young children’s learning in exciting, new ways. We want to ensure that the robot is a polite and socially-inclusive playmate—so we will be developing new capabilities for the robot to interact with small groups (e.g., adult-child-robot or child-child-robot). While the robot is not designed to replace or compete with teachers, it should serve as a fun and engaging practice partner especially for children who need additional support. As part of this program, we will be collecting speech data from young children (strictly with their parents’ permission) so that the robot can improve its understanding of what young children say. Teachers will be able to track children’s progress with the robot and the feedback the robot provides to the teachers can be very useful and helpful. Throughout our process, we will be iterating with our co-designing teachers using their input to ensure a useful and easy-to-use robot design . At Dubai Heights Academy (DHA), the robots will first be available in an afterschool setting. A member from our MIT research team will support DHA staff and students onsite with the robot. We are also working with pre-K and kindergarten classrooms in the U.S. to better understand how social robots enhance and augment classroom time. We are exploring home use to understand how social robots can help reinforce learning at home with parents, siblings and caregivers. Especially during this time of remote learning for many young children, we’re just beginning to understand how social robots may be a critical technology for early childhood education when in-school learning is not possible. Toward the end of our three-year exploration, we expect to have developed a social robot learning experience that can be piloted and evaluated for learning efficacy in the next phase of the project.
2020-2021 school year at Dubai Heights Academy
This academic year, beginning later in the first term of 2020, we are introducing Y6 and Y7 students at Dubai Heights Academy to learning modules on computational thinking. Specifically, we are bringing the CoolThink curriculum to DHA students. CoolThink was developed jointly over the past three years by MIT, the Education University in Hong Kong, the City University of Hong Kong and the Hong Kong Jockey Club and has been rigorously evaluated by the Stanford Research Institute.
The CoolThink curriculum is currently used in 32 Hong Kong upper-level primary schools and is being scaled to include 150 schools. Our goal this year for DHA is to prepare students with a solid foundation for our Y7 — Y9 AI literacy curriculum. Over time, we will deepen our offering for Y7 to include more AI Literacy modules while we also expand the curriculum to Y8 and Y9. We anticipate students will do these activities on average one session per week during students’ Computer Science class. Sessions will be led by the DHA Computer Science teacher and a MIT mentor. The social robot learning companion work will begin next year beginning in the first term of 2021.
Cynthia Breazeal (PI)
MIT Media Lab
Eric Klopfer (Co-PI)
MIT STEP Lab
Hae Won Park MIT
Hal Abelson (Co-PI)
MIT STEP Lab
MIT STEP Lab
Glenda Stump (Co-PI)
MIT Open Learning
MIT STEP Lab
MIT STEP Lab
Mary Cate Gustafson-Quiett
MIT STEP Lab