Will Artificial Intelligence make educators obsolete
Artificial intelligence and machine learning have been making headlines for decades, often with fanciful claims about what is around the corner. However, the New Directions Conference held in Shanghai last year, which saw more than 60 speakers deliver papers to over 500 delegates, provided some solid insights into the current state of the field and what to expect in the near future for language learning and assessment, and whether technology will make educators obsolete.
Among the topics discussed were language proficiency frameworks, foreign language testing, measuring teachers' language teaching proficiency, assessment literacy for language teachers, and artificial intelligence and technology. In terms of AI and technology, the conference covered topics ranging from automated language assessment to computer course text selection to the future role of technology in education.
Co-founder and Chief Scientist of Liulishuo (LLS) Hui Lin, highlighted the advantages of AI in learning and assessment, explaining it has the potential to provide effective solutions at affordable prices: “Because AI is scalable it is very affordable and doesn’t have the same limitations of human services. Previously, there was no product available in this price range.” Founded in 2012, LLS is a new-tech startup which has developed an online language testing app in China that claims to provide instant feedback on speaking test performance. The company currently boasts 50 million users, with 1 million subscribing to their paid service.
Lin, who previously worked for Google on speech recognition, is confident about the future. “Since the launch of our paid service in 2015, we’ve seen 50 percent growth quarter on quarter.” Hui goes on to explain how the system’s algorithm was developed through mass sampling analysis and hopes to be able to calibrate the test more accurately to reflect results from established, traditional testing systems. Hui very much sees the future of assessment moving toward automation.
But as Alistair Van Moere, Chief Product Officer for MetaMatrics, points out, AI also has huge potential in assisting teachers in a conventional classroom environment. MetaMetrics, for example, is attempting to bring meaning to measurement by matching students to resources with scientific, universal scales. According to Van Moere, MetaMetrics has used AI to gauge the complexity of texts in order for teachers and learners to find texts appropriate for their learning level. Van Moere explained how the company was able to use machine learning to develop the Lexile Scale – a computer system that is capable of judging the complexity of a text. Their database provides Lexile measures for nearly 300,000 books, and their Lexile Analyzer is available for educators to rate individual texts for use in various educational environments; even Amazon relies on the service to aid customers in their reading selection.
Despite the astonishing advances in the field of AI and machine learning, the conference also identified some of the difficulties ahead. In the final panel discussion, Wei Si of iFLYTEK explained the complexities involved in creating learning algorithms for evaluating written language, pointing out that despite the enormous progress already achieved in the field in terms of identifying grammar structures and complexity and accuracy of language usage, AI is still a long way from something we all take for granted – grasping innate or intuitive meaning. Wei provided the following statement as an example: “He cannot lift his son anymore because he is too heavy.” Programming a machine to tell who is too heavy is a still a long way away.
So perhaps technology does not spell the end of the educator – at least not yet. Maybe, as one participant put it, technology will change the role of educator from “the sage on the stage” to the “guide by the side”.