Watson and Elmo are joining forces.
Computing giant IBM and the nonprofit organization behind popular children’s television show “Sesame Street” announced Wednesday a new partnership to bring so-called “cognitive computing” to early childhood education. The idea is to develop products featuring Sesame characters and content that are capable of responding in personalized ways to individual students based on their skills, interests, and developmental trajectories. Key to the effort will be the technology behind IBM’s Watson, which takes a “big data” approach to understanding and responding to human language.
“We’re in the ‘anything is possible’ research-and-development stage,” said Steve Youngwood, Sesame Workshop chief operating officer, in an interview.
“If we can take our learning and creative expertise, and combine that with IBM’s [technological] expertise, we hope to come out with products that are very personalized for kids and teachers and caregivers.”
Possible applications include apps and toys that are capable of processing children’s natural language to detect patterns in their vocabulary or speech development, for example. Watson would then compare the resulting information against its massive store of data on human language development, enabling the app or toy to identify the response that would best help that individual child learn. Other possibilities include reading apps that customize content to individual students’ interests and skill levels and tools that provide teachers with better information about each student’s strengths and weaknesses, Youngwood said.
Technologies that adapt to individual users in limited ways have long been available, even for young children.
But Watson and its “cognitive computing” approach is different, said Satya Nitta, IBM’s program director of cognitive sciences and education technology, in an interview. Rather than following a decision tree made up of “if-then” type statements that have been programmed in advance, Nitta said, Watson is able to search for patterns among massive amounts of unstructured data, identify trends and categories on the fly, and vary its responses based on what it “learns.”
“Once you marry that with the ability to interpret natural language, things become really interesting,” Nitta said. “If Watson poses a question to a student, and the student answers in their own words, the computer is able to gauge whether they get the concept or not, if the likelihood is high or low, and what else they are likely to be struggling with.”
Early applications of the technology are currently in the prototype stage, with plans to pilot some products in classrooms and possibly begin bringing tools to market as soon as the end of 2016. Long term, the partnership could result in a platform that would be open to other educators and entrepreneurs seeking to develop personalized learning tools for young children.
Watson first came to public prominence in 2011, when it successfully defeated human experts on the popular television game show “Jeopardy!” The technology is already in use in other sectors, including healthcare, where Watson is helping doctors to more effectively diagnose and respond to the needs of patients, according to Stanley Litow, IBM’s vice president of corporate citizenship and corporate affairs.
The new announcement comes at a time of increased focus on the idea of bringing “personalized learning” into K-12 and higher education. Philanthropists such as Microsoft’s Bill Gates and Facebook’s Mark Zuckerberg, for example, have committed huge sums to help spur the development of new technologies and school models aimed at providing students with a more customized educational experience.
But the research base around the impact of technology-enabled personalized learning on student outcomes remains thin, at best.
Watson vs. Human Interaction?
And the move to target such approaches at infants, toddlers, and preschoolers is sure to cause some unease among parents and early childhood experts, many of whom remain focused on the importance of high-quality human interactions between young children and adults as the key to language acquisition, emotional and social growth, and brain development.
“It’s always going to be contingent on when and how [the technology] is used,” said Kathy Hirsch-Pasek, a psychology professor at Temple University in Philadelphia. “Little people really like and benefit from human interaction. While we can mimic that digitally, it’s not the same thing.”
On one hand, Hirsch-Pasek expressed hope that the new IBM-Sesame partnership might lead to the development of products and tools capable of responding in ways that are “socially responsive” to children’s own language, attention, and actions. She gave the example of Skyping with her 1-year-old granddaughter while using puppets: When the girl looks at a ball and verbalizes, Hirsch-Pasek said, she is able to direct the puppets to immediately say “That’s right! It’s a ball,” following the child’s lead and responding in a way that is timely, relevant to the child’s own intent, and helpful in advancing the child’s ability and understanding.
“There’s a whole literature to suggest that children learn language when you follow them,” Hirsch-Pasek said. “But that kind of ‘adaptive contingency’ is something the technology hasn’t cracked yet.”
Much less exciting, however, would be interactive technologies focused on delivering customized content information, Hirsch-Pasek said.
“Don’t make an Elmo doll that is going to spit forth knowledge to somehow fill the empty head of a child,” she said.
Data Privacy, Security Questions
Also a potential concern: data privacy and security.
Nitta and Litow of IBM said that many of the products under consideration are not dependent on collecting large quantities of data from an individual child or school, although they acknowledged that such collection will be “extremely helpful” to development of the technology. The intent is to make such data collection opt-in, and both IBM and Sesame Workshop are committed to implementing appropriate safeguards, officials said.
Ultimately, it seems likely that the groups will hope to strike a bargain with parents and schools in which it is allowed to collect and store data so long as such information is anonymized (to limit the possibility of being tied back to a specific child) and used only to improve educational products and services. The K-12 sector is currently working through similar issues, with both district-vendor contracts and state laws evolving to reflect various concerns and tensions.
Youngwood of Sesame Workshop stressed that their work is in its early stages, with much yet to be worked out.
The new partnership is ripe with both potential and pitfalls, said Temple’s Hirsch-Pasek.
“This is absolutely where we need to go next, and if they can do this, it will make for wonderful advances in our understanding of human social behavior and learning,” she said.
“I very much hope that Sesame Workshop will continue to be a trusted source for all of us.”
Photo: “Sesame Street” puppets are seen in Hamburg, northern Germany, Jan. 22, 2003. —Christof Stache/AP-File
A version of this news article first appeared in the Digital Education blog.