Sydney Students Allegedly Forced to Retake NAPLAN After AI Predictive Text Error

A predictive text malfunction allegedly compromised the integrity of the NAPLAN writing exam at two Sydney schools, Waverley College and Kambala, in that it allowed students to access AI-driven text suggestions. Caused by a technical oversight, this incident led to affected students being required to retake the test.

Risk classification

• Primary risk domain: 7 AI system safety, failures, & limitations
• Primary risk subdomain: 7.3 Lack of capability or robustness

The incident involved a technical failure where the testing environment or device configuration failed to robustly restrict the AI predictive text feature during a high-stakes exam.

Additional risk subdomains

• 5.1 Overreliance and unsafe use: The active AI feature presented an unauthorized and unsafe assistance capability in a critical educational assessment context.

Causal factors

• Entity: Other
• Intent: Unintentional
• Timing: Post-deployment

The incident was caused by a technical glitch or administrative oversight that unintentionally left the predictive text feature enabled during a standardized exam.

EU AI Act risk tier

• Risk tier: 4 Minimal or No Risk

Minimal or No Risk: Apple's predictive text is a general-purpose writing assistant tool that poses low or negligible risk to users in normal operation, even though its unauthorized presence in an exam created administrative complications.

AI system and alleged parties

• AI system: Apple predictive text (Apple)
• AI purpose: Predictive Typing; Writing Assistant
• Behaviour type: Assistant
• Alleged developer: Apple
• Alleged deployer: Waverley College (Sydney, Australia), NSW Education Standards Authority (NESA), Kambala (Sydney, Australia), Australian Curriculum, Assessment and Reporting Authority (ACARA)
• Alleged harmed parties: Waverley College (Sydney, Australia) teachers, Waverley College (Sydney, Australia) students, Kambala (Sydney, Australia) teachers, Kambala (Sydney, Australia) students

Harm severity

Highest direct severity in any category: Negligible. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

People affected

• Occurrences reported: 2
• People reportedly harmed: 150
• People reportedly exposed: 150

Potential causes

Management

• Inadequate Technical Oversight: School IT management did not verify that AI features were fully disabled.

Technology

• Active AI Predictive Text: AI-powered predictive text remained active on student devices during test.

Human Factors

• Failure to Disable AI Features: Users and staff failed to manually turn off the AI writing assistant.

Process and Methods

• Inadequate Device Preparation: Instructions to disable AI predictive text were not successfully executed.

Regulatory Environment

• Lack of Automated Enforcement: Regulators lacked automated tools to block AI features on student devices.

Information quality

• Classification confidence: High
• Reason for confidence: The reports clearly describe the incident, the schools involved, the specific AI feature (Apple predictive text), and the resolution (resitting the exam). There is no ambiguity about what occurred, though exact student numbers must be estimated.
• Ambiguities identified: The exact number of affected Year 5 students is not specified, and it is mentioned that other schools might have been affected without realizing.
• Alternative interpretations: None.

An unintentional administrative oversight left Apple's predictive text feature enabled during Australian NAPLAN writing exams at two schools. While requiring a test retake, the incident has negligible national security implications and represents a localized administrative issue rather than a threat to national infrastructure, sovereignty, or economic stability.

• Overall national security impact: Negligible
• Response level: Minor
• Scope: Single nation
• Primary target: Australia
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. No active national security crisis; localized administrative issue already resolved.
• Autonomy: Human-controlled. Predictive text acts as a typing assistant requiring user input and final text selection.
• Novelty: Established threat. Administrative oversights and software configuration issues during standardized testing are established issues.

Impact by dimension

• Physical security: Negligible. No physical security or critical infrastructure threat; limited to high-stakes standardized school testing.
• Information security: Negligible. No intelligence compromise, classified model theft, or systematic information warfare indicated.
• Sovereignty: Negligible. Administrative glitch in school testing does not impact core government functions, sovereignty, or electoral processes.
• Economic security: Negligible. No threat to economic stability, strategic industries, or national technological competitive advantage.
• Societal stability: Negligible. School exam rescheduling does not impact societal stability, civil liberties, or broader population safety.