This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Nevada implemented an AI system developed by Infinite Campus to identify at-risk students for funding purposes. The model significantly reduced the number of students classified as at-risk, leading to substantial budget cuts for schools serving low-income populations and raising concerns about transparency, algorithmic bias, and the adequacy of support for vulnerable students.
An AI system developed by Infinite Campus and deployed by Nevada to identify at-risk students led to a sharp reduction in the number classified as needing support, dropping from 270,000 to 65,000. The reclassification caused significant budget cuts in schools serving low-income populations. The drastic reduction in identified at-risk students reportedly left thousands of vulnerable children without resources and support.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Unequal treatment of individuals or groups by AI, often based on race, gender, or other sensitive characteristics, resulting in unfair outcomes and unfair representation of those groups.
Also classified as
The system failed to robustly identify clearly vulnerable students, such as homeless and low-income children, as being at risk.
The algorithm's specific weights and decision-making processes were kept proprietary by Infinite Campus, leaving educators frustrated by the 'black box' system.
The AI system initially utilized sensitive demographic characteristics like gender, race, and birth country, leading to concerns of unfair discrimination where students with identical academic profiles received different risk scores based on gender.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The budget shortfalls and exclusion of vulnerable students were unexpected outcomes resulting from the deployment of the AI model to optimize funding allocation.
The deployment of an AI-based student risk model in Nevada led to significant, biased reductions in educational funding for over 200,000 vulnerable students. While causing notable domestic policy challenges and highlighting risks in algorithmic governance, the incident presents minimal threats to core national security.
The AI system acts as an Assistant by analyzing student data to predict graduation risks and outputting a 'grad score' to help school administrators allocate funding, though the state chose to rely solely on these scores for its funding decisions.
Risk Level 2: High Risk. The system is an educational tool used to determine student risk levels and allocate funding, which directly affects access to education. The EU AI Act classifies 'Educational and vocational training systems affecting access to education' as High Risk.
The primary goal of the AI system was to identify which students were at risk of falling behind to help the state allocate educational funding resources. It also served to identify specific students matching these risk profiles.
The report provides clear details on the AI developer (Infinite Campus), the specific metrics used, the scale of the impact (reducing at-risk students from 270,000 to 65,000), and the resulting funding issues for specific schools. The role of the AI is explicitly defined.
The funding cuts could be viewed as a policy decision by the state of Nevada on how to define 'at-risk' rather than purely an AI failure, though the AI's specific classifications directly drove the outcome.
Technology
Proprietary Black Box Model
The algorithm workings were kept private as proprietary intellectual property.
Narrow Definition of Risk
Model focused solely on graduation rates, ignoring mental health and wellbeing.
Algorithmic Bias in Variables
First-year model used sensitive factors like race, gender, and birth country.
Data Inputs
Proxy Metrics for Engagement
Using guardian portal logins as a proxy for student risk is flawed.
Incomplete Risk Indicators
Data inputs lacked metrics on student depression, self-harm, and food security.
Human Factors
Overreliance on Predictive Tech
State administrators trusted the model output without local verification.
Lack of Technical Literacy
School leaders did not understand how the automated scoring system worked.
Process and Methods
Lack of Transparency
No clear explanation was provided to schools on how scores were calculated.
No Validation Mechanisms
The state implemented the system without independent auditing or evaluation.
Abrupt Funding Reallocation
Budgets were slashed suddenly based on model outputs without transition plans.
Management
Prioritizing Cost Efficiency
Management sought to target dwindling funds without assessing negative impacts.
Inadequate Risk Assessment
Decision-makers failed to foresee the impact of dropping 200,000 students.
Regulatory Environment
No Algorithmic Oversight
There were no state regulations governing the use of AI in school funding.
Absence of Standardized Audits
No regulatory requirement existed to audit the predictive model for bias.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
