New Zealand Schools Deploy Google Gemini for Student Assessment as Education Ministry Reviews AI Policies

New Zealand schools are trialling Google Gemini for automated student assessment and feedback, with early results showing 40% faster marking times but raising concerns about AI bias in educational evaluation. The Ministry of Education is now accelerating its review of artificial intelligence governance frameworks following mixed reactions from teachers and parents.

Several secondary schools across Auckland and Wellington have quietly begun integrating Google Gemini into their assessment workflows over the past month, marking a significant shift in how New Zealand educators approach student evaluation. The AI system is being used primarily for initial essay scoring, grammar correction, and providing structured feedback on written assignments across English, history, and social studies curricula.

Mount Albert Grammar School reported that Google Gemini reduced their English department’s marking workload by approximately 35 hours per week, allowing teachers to focus more on personalised instruction and curriculum development. However, the implementation has not been without controversy, as some educators question whether artificial intelligence can adequately assess the nuanced thinking and creativity that characterise quality student work.

The trial results have revealed both promising efficiency gains and concerning inconsistencies. Google Gemini demonstrated particular strength in identifying grammatical errors, structural issues, and factual inaccuracies in student essays. Yet teachers noted the system occasionally marked down creative writing that deviated from conventional structures, potentially penalising innovative thinking that human assessors would recognise as advanced.

Wellington College’s deputy principal Sarah Henderson expressed cautious optimism about the technology while acknowledging its limitations. “Gemini excels at the technical aspects of assessment, but we’ve observed instances where it fails to recognise sophisticated arguments that don’t follow traditional essay formats,” Henderson noted. “Our concern is that students might begin writing to satisfy the AI rather than developing authentic voice and critical thinking skills.”

The deployment coincides with mounting pressure on New Zealand’s education system to modernise assessment practices while maintaining academic rigour. According to Ministry of Education, the agency is developing comprehensive artificial intelligence guidelines expected to be released by mid-2026, addressing both opportunities and risks associated with AI integration in classrooms.

Google’s education division has positioned Gemini as a teaching assistant rather than a replacement for human judgment, emphasising that final assessment decisions should remain with qualified educators. The company’s New Zealand education lead, Michael Chen, argued that Gemini serves as a “first-pass filter” that identifies obvious errors and provides consistent baseline feedback, freeing teachers to focus on higher-level pedagogical concerns.

Yet this narrative faces scrutiny from educational researchers who warn about the potential for algorithmic bias to influence student outcomes disproportionately. Dr. Lisa Patel from the University of Auckland’s Faculty of Education highlighted concerns about AI systems inadvertently discriminating against students from diverse cultural backgrounds whose writing styles may not align with the training data used to develop these models.

The economic implications extend beyond individual schools, with the Education Review Office estimating that widespread AI adoption could reduce marking costs across New Zealand’s secondary education system by up to $12 million annually. This potential saving comes at a time when schools face increasing financial pressures and teacher shortages, making automation an attractive proposition for stretched education budgets.

Parent reactions have been decidedly mixed, with some welcoming faster feedback turnaround times while others express concern about their children’s work being evaluated by artificial intelligence. The Auckland Secondary Schools Parents Council has called for greater transparency about when and how AI is being used in assessment, arguing that parents have a right to understand the tools shaping their children’s educational outcomes.

Industry observers note parallels with the introduction of plagiarism detection software in the early 2000s, which initially faced resistance before becoming standard practice in New Zealand schools. However, the stakes appear higher with Google Gemini, as the system doesn’t merely detect academic dishonesty but actively participates in the assessment process itself.

The broader implications for New Zealand’s education technology sector are substantial, with local edtech companies now scrambling to develop competing AI assessment tools or partnership arrangements with international providers. This rush toward automation raises questions about data sovereignty and whether student assessment data should be processed by overseas technology companies, regardless of their reputation or capabilities.

As the Ministry of Education prepares its forthcoming guidelines, the current school trials serve as a crucial test case for balancing innovation with educational integrity. The challenge lies not in whether artificial intelligence can improve assessment efficiency, but whether it can do so without compromising the human elements of teaching and learning that define quality education. The next six months will likely determine whether Google Gemini becomes a standard tool in New Zealand classrooms or remains a cautionary tale about premature AI adoption in education.