Fright night....Why Schools and Colleges remain fearful of AI

Barriers to AI and Technology Adoption

While AI and technology offer significant benefits, schools and educational organisations still remain reluctant due to various challenges.

The UTAUT model (Venkatesh, 2003) remains a great resource when exploring what deters schools from responding to and experimenting with AI.

Understanding Technology Acceptance

Adopting technology like AI requires understanding key factors that influence decision-making. The Unified Theory of Acceptance and Use of Technology (UTAUT) provides a helpful framework for identifying and addressing these factors:

• Performance Expectancy: Show how AI can improve productivity and performance by offering tangible benefits!

• Effort Expectancy: Reduce the perceived complexity of AI by providing thorough and sustained training and support.

• Social Influence: Highlight success within the school through case studies and embed this within CPD.

• Facilitating Conditions: Ensure that organisational support is in place, such as offering on-demand training and promoting the use of ‘knowledgeable others’.

AI has significantly accelerated the need for users to try, test, and experiment with new platforms. However, misconceptions persist, such as the belief that large language models (LLMs) are "thinking" like humans. Despite AI's growing presence, many of the traditional barriers to technology acceptance—such as concerns about trust, complexity, and potential biases—remain.

1. Individual Concerns

Barrier: Trust Issues

Many teachers and senior leaders are hesitant to trust AI, fearing inaccuracies or biases in AI outputs, especially when the system’s inner workings are unclear. i.e. how the hell does this sorcery work!

Solution:

• Understanding: It's essential school leaders understand how AI tools work, including how data is processed and decisions are made.

Barrier: Time and Skill Requirements

Learning and effectively using AI tools can feel overwhelming, especially with the myriad of other responsibilities which teachers and leaders face.

Solution:

• Offer Simplified Training: Provide step-by-step guides and tutorials that focus on the essentials. Deliberate and sustained training opportunities should be built into core CPD alongside other priorities.

• Ongoing Support: Create a support system, such as steering groups or knowledgeable others, where users can ask questions and share experiences as they learn.

Barrier: Bias Concerns

AI systems can unintentionally reflect biases found in training data, leading to unfair outcomes, particularly when generating lesson content or activities.

Solution:

• Challenge the LLM: Ensure your prompts include challenges to AI, asking it to represent diverse and contextualised considerations.

• Be aware: Humans are not unbiased! Be aware that those who design, build, and curate the LLMs are likely to do so with their own views and perspectives baked in.

  
  

2. Technology-Related Obstacles

Barrier: Data Privacy and Security

Concerns about how personal data is used, stored, and protected can make senior leaders reluctant to adopt AI.

Solution:

• Clear Privacy Policies: Clearly communicate how data is collected, stored, and used.

• Avoid using sensitive data in LLMs where possible: Despite new additions where users can choose to prevent LLMs from training on their data, they remain prone to the same security handicap as many platforms.

Barrier: Quality and Reliability of AI Outputs

Many teachers may doubt the reliability of AI, especially if they’ve encountered inaccurate or misleading outputs. Especially those who are not yet skilled in the art of prompting!

Solution:

• Iterate: Rarely do LLMs produce fantastic responses initially. Refining, amending, and curating the response often leads to much better outcomes.

• Be clear: ‘Prompt engineering’ became the new term of 2024! However, a PhD in AI is not needed to get the most from your LLMs. Being clear, specific, and building the context in which the LLM should respond often leads to high-quality outcomes. Dan Fitzpatrick has an 8-levers approach well worth using.

  
  

3. Business-Related Barriers

Barrier: Cost of Implementation

The upfront costs of adopting AI, such as software purchases, hardware upgrades, and staff training, can be a deterrent for schools in a frugal environment. Many LLMs are available free of charge but have limitations bound by their subscription levels.

Solution:

• Pilot Programmes: Start small with free LLM licenses or involve a small series of individuals in subscriptions before diving headfirst into team licenses! Approaching this on a monthly basis and evaluating the benefit is a great place to start.

Barrier: Compliance with Regulations

Navigating the complex and evolving landscape of AI regulations can be challenging for many in schools who are time-poor.

Solution:

• Stay Informed through L&D: Prepare staff with thorough and robust training on the regulations behind using AI responsibly. The EU AI Act (2024) mandates AI literacy training for employees of companies using AI, including schools, from February 5, 2025!

Barrier: Organisational Resistance

Teachers and senior leaders alike may resist AI adoption due to concerns about job displacement, trust issues, or the perceived complexity of AI.

Solution:

• Be open: Be transparent about AI’s role within the school, emphasising how AI can enhance jobs rather than replace them. It’s a tool, much like being able to use Excel!

• Upskilling Opportunities: Offer training that empowers teachers to work alongside AI, highlighting how it can improve their roles and job satisfaction. Base this on tangible benefits such as boosting lesson creativity, differentiation support, and time-saving exercises.