Nevada AI student risk model prompts funding controversy

An AI-driven system introduced by Nevada to assess which students are at risk of academic failure sparked controversy about its effectiveness, impact on school funding, and downstream impacts on students' health and wellbeing. What happened Developed by Infinite Campus and introduced early 2024 by Nevada's Department of Education, t he "G RAD scores" algorithm evaluates over 75 factors, including academic perform ance, attendance and disciplinary records, demographic , income and "family engagement " metrics. Each student receives a grad score ranging from 50 to 150, where a lower score indicates a higher risk of not graduating. The model is reported to have dramatically reduced the number of students classified as "at risk" from over 270,000 to less than 65,000. However, this has also led to substantial cuts in state funding for schools that relied on these classifications for financial support. It also led to questions about whether the system is fair, and led experts to question whether the system's emphasis on graduation rates may be too narrow, neglecting other critical aspects of student well-being. Educators have noted an increase in challenges related to mental health and self-harm among students who might not be classified as at risk under the new criteria. Why it happened Nevada aimed to improve its historically uneven school funding system, which had been widely criticised for providing significantly less financial support to low-income districts. Low-income districts there have nearly 35 percent less money to spend per pupil than wealthier ones do - the largest gap of any US state. By using AI to refine the criteria for identifying at-risk students, Nevada sought a more targeted approach and set a much higher bar. In addition, Infinite Campus refuses to disclose how its model works, claiming it is proprietary, prompting scepticism about the fairness and transparency of the system and its governance. What it means Nevada's AI system may improve efficiency, but it may also be unfair and overlook the broader needs of low-income students, resulting in even greater disparities and a welter of downstream impacts including increased student mental health issues and state healthcare costs. The cuts driven by the system are also seen to serve as a possible justification for future cuts to educational funding across the state. What we think As a public institution, Nevada's education authority should not use a system it appears not to fully understand and for which details of its inner workings are not made available. Claiming a system is proprietary when it is used in the public domain is not a legitimate or ethical argument, and Infinite Campus should disclose the full details of its system, notably its model weights, without delay. System 🤖 GRAD Scores 🔗 Developer: Infinite Campus Country: USA Sector: Education Purpose: Predict stud ent graduation likelihood Technology: Machine learning; Prediction algorithm Issue: Accuracy/reliability; Fairness

Date it happened

2024-02-01

Organisation involved

Nevada Department of Education

Product, system or model

GRAD Scores

This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.