AI Learning Analytics Monitor: Emotion Recognition Data Packaged as Precision Teaching is Actually a Targeted Sales Scam

Who Gets Targeted

The victims are primary and secondary school management and anxious student parents. School representatives are mostly informatization directors eager to achieve political accomplishments using AI labels, yet lacking judgment on algorithm principles and data compliance, making them easily swayed by demo presentations and free trials. Meanwhile, parent groups panic as their children are labeled with learning difficulties or attention deficits, and under the pressure of targeted anxiety-inducing sales scripts from manufacturers, they purchase thousands of yuan worth of tutoring courses or learning devices. The common psychological vulnerabilities of these groups include a blind belief in AI technology authority, a superstition about quantified data, and educational anxiety of not wanting to lose at the starting line, causing them to unquestioningly purchase system-generated learning reports.

骗局怎么运作

• Step 1: Approach schools using sales pitches of 'free use' and 'proven results,' promising smart classroom pilots and national research project cooperation to lower the school's decision-making threshold. The actual goal is to use the school's public credibility to legally enter classrooms and batch-collect student data, often claiming that a neighboring district's school has already used it for a semester with significant grade improvements.
• Step 2: Deploy cameras and tablet terminals in classrooms. The system labels student attention, smile frequency, head-down frequency, and even frown frequency in real-time, disguising simple threshold rules as AI emotion recognition algorithms to create an advanced technological illusion, while lacking scientifically validated models underneath.
• Step 3: Package the collected behavioral data into personalized learning profiles, generating radar charts and weakness reports for each student while intentionally exaggerating problem items, such as judging excessive frowning as a learning difficulty, to create excuses for subsequent commercial conversion.
• Step 4: Cross-platform data integration. The platform shares tags with off-campus tutoring institutions, and pushes high-priced tutoring courses and learning hardware to parents of students judged as struggling learners. Parents pay out of fear of their children falling behind, and the platform takes a commission, completing the closed-loop from public-interest campus entry to traffic monetization.
• Step 5: When faced with questioning, intellectual property and algorithm confidentiality are used to refuse the disclosure of rationale. Once exposed, the company entity and brand name are changed to re-bid. The same system continues to be replicated in the next district under a new name, making the cost of rights protection far lower than the cost of the company running away.

红旗信号（看到这些快跑）

• 🚩 The system claims to judge comprehension ability using facial expression recognition, yet cannot provide third-party educational psychology assessment certification or algorithm filing numbers.
• 🚩 The so-called personalized report templates are highly standardized, with weakness descriptions for different students being highly similar, differing only in data chart colors.
• 🚩 The platform requires the collection of sensitive student information such as facial videos and classroom speech recordings, but has not signed clear informed consent forms with parents.
• 🚩 Following the generation of learning reports, associated tutoring courses and learning machine purchase links are intensively pushed, turning the report itself into a marketing funnel entry point.
• 🚩 The company has a short establishment history, frequent name changes, inconsistent bidding and operating entities, and refuses to provide intellectual property proof for question banks and algorithms.
• 🚩 It enters campuses under the guise of free pilots and project research, but buries vague clauses regarding data authorization and commercial conversion in the contract.

真实案例

• On May 12, 2026, Qilu Evening News page A05 reported that an educational platform labeled student attention, smile frequency, and head-down frequency in classrooms in real-time while synchronizing with a learning database. After judging excessive frowning as a learning difficulty, it targeted parents with high-priced tutoring courses. Journalist investigations also found cross-platform data integration issues, with involved personnel interviewed under pseudonyms.
• On June 2, 2026, the Ministry of Education issued a gaokao early warning, naming and warning against false propaganda methods such as AI answer-prediction, pointing out that some institutions use AI concepts to package predictions and score-improvement promises for profit, reminding examinees and parents to beware of scams, reflecting that false propaganda in AI education has become a regulatory focus. (Source: [https://education.news.cn/20260602/ed33b6d64d374b4f88c2cea92db34984/c.html](https://education.news.cn/20260602/ed33b6d64d374b4f88c2cea92db34984/c.html))
• In February 2026, the OECD warned in its 'Digital Education Outlook 2026' report that student reliance on AI tools creates an illusion of knowledge mastery, and that so-called personalized recommendations often fail to genuinely improve learning outcomes, providing an international authoritative reference for identifying pseudo-adaptive learning systems. (Source: [https://www.163.com/dy/article/KKMOTTH60514BQ68.html](https://www.163.com/dy/article/KKMOTTH60514BQ68.html))

Official Stance

• On June 2, 2026, the Ministry of Education issued a gaokao early warning, explicitly reminding the public to watch out for false promotional tactics such as celebrity teacher predictions and AI predictions, and to guard against scams.
• On July 6, 2026, the Cyberspace Administration of China reported on the first phase of the special action to clean up chaos in AI application governance, focusing on chaos such as failing to fulfill large model filing obligations and AI data poisoning, and opening a reporting hotline area.
• In April 2026, the Cyberspace Administration of China, together with multiple departments, issued the 'Interim Provisions on the Administration of Artificial Intelligence Anthropomorphic Interactive Services,' and five departments including the Ministry of Education issued documents related to AI plus education to standardize AI education applications.
• In February 2026, the OECD's 'Digital Education Outlook 2026' report warned of the illusion of knowledge mastery risks in AI education applications, calling for a prudent evaluation of the claimed effects of learning AI products.

How to Protect Yourself

• ✅ Before school procurement, suppliers are required to present algorithm filing numbers, third-party educational evaluation certifications, and data security assessment reports. Anyone refusing disclosure is strictly excluded.
• ✅ Before signing contracts, legal counsel reviews data collection clauses, ensuring that sensitive student information such as faces and voices must not be used for commercial marketing, and separate parental informed consent is required.
• ✅ Set up exit mechanisms and data deletion clauses during the pilot period, requiring vendors to issue data destruction certificates and retain audit rights after the pilot ends.
• ✅ When parents receive course recommendations attached to learning reports, they should first verify with head teachers and public teaching research departments, guarding against the combined marketing of reports plus course sales.
• ✅ Upon discovering clues of data abuse or false propaganda, report and retain evidence simultaneously through the Cyberspace Administration of China's dedicated AI application chaos reporting channel and the 12315 consumer hotline.