What are Intel's plans, uses and criticisms of AI that analyzes students' lesson attitudes?

It is difficult to understand how much a student understands what he / she is learning in a school lesson, whether he / she is able to concentrate on the lesson, and whether he / she has any troubles, etc., unless he / she directly reveals it. Classroom Technologies , a company that sells lesson software in virtual space, is developing a system in partnership with Intel that analyzes students' facial expressions and how they approach teaching materials to help them understand their lesson attitudes and understanding. Can be detected. While it is expected to provide fairly useful information for educators, there are also numerous criticisms that value student rights.

Class tests Intel AI to monitor student emotions on Zoom --Protocol
https://www.protocol.com/enterprise/emotion-ai-school-intel-edutech

Intel is developing a technology that uses artificial intelligence (AI) and image / video analysis to read the state of a student's face capture and learning efforts. Intel researcher Cinem Aslan , who was involved in the development, said, 'With this system, teachers recognize when students need help, provide teaching materials that are suitable for students, and help students not get bored in class. It will enable large-scale tutoring. '

'By introducing Intel's Educational Analysis AI into remote lessons in virtual space, we can improve communication between teachers and students, which is difficult without face-to-face,' said Michael Chasen, a collaborator at Classroom Technologies. I can do it, 'he says, saying his expectations for AI systems in the field of education.

On the other hand, there are two major criticisms about analyzing students' facial expressions and attitudes with AI. The first is the feasibility that 'it is impossible to accurately judge emotions such as boredom, confusion, happiness and sadness from facial expressions and other external information in the first place'. Kate Crawford, who studies AI social impact, criticized making false guesses about abilities and emotional states from human appearance, calling it 'Phrenological Impulse' and 'even with the same facial expression. Reading emotions by ignoring the context and using only facial movements can be misleading. '

Experts warn of the danger of 'emotion recognition AI', what is the problem? --GIGAZINE

There are also criticisms from a moral point of view, such as the anxiety that facial recognition monitors students when verifying the identity of individual students for the purpose of security and prevention of fraudulent examinations, and the significant invasion of privacy. increase. In addition, since face recognition technology tends to have a high false positive rate for colored races, it is often argued that face recognition technology is banned because it is racially discriminatory.

Big cities in the United States ban face recognition systems one after another --GIGAZINE

by Mike MacKenzie

'We are sensitive to people's personal rights and we need to prevent excessive intrusion into their rights by the system,' Chason said, with a good understanding of the criticisms of facial recognition AI. .. Also, Aslan is better off teaching teachers to ensure correct and appropriate lessons and when students are asking for help, rather than using Intel technology to monitor or detect and punish students for fraud. He says the prospect is to provide additional information.

But on the other hand, Nandia Sampas, an analyst on AI bias and accountability issues in Consumer Reports , a non-profit consumer organization, said, 'With Intel's facial recognition AI technology, it's unintentional. But it can monitor and punish students. '

Introducing Intel's facial recognition AI technology into Classroom Technologies' virtual lecture software is in the testing stage, and Chason said, 'Whether the results of AI analysis are related to student performance, useful data in education from there. 'We need to make sure we can get it,' he said, 'I don't think it's a fully mature technology yet. The data analyzed is just an idea and used alone without context.' You shouldn't do it. '