Appendix

Appendix 1: CEDoW Endorsed Gen AI tools

Google Gemini

Access from the “9 dots” in Google or https://gemini.google.com/

Google Gems

Access within Gemini or https://gemini.google.com/

Notebook LM

Access from the “9 dots” in Google or https://notebooklm.google.com/

Co Pilot (via CEDoW login)

Use caution to ensure you are using CEDoW login via Microsoft

https://copilot.microsoft.com/

CEnet’s CeChat

https://cogniti.cenet.ai/

If any staff member/school wishes to have a specific Gen AI tool endorsed by CEDoW, or if you would like advice about whether to use a particular Gen AI tool, please contact one of the Learning Technology staff within the Improvement, Learning and Wellbeing Team.

Learning about specific Gen AI tools/platforms

Staff will benefit from a deeper understanding of how CEDoW endorsed Gen AI tools can be used safely to support their work. Some specific AI Insights on the following tools/platforms are planned in the short term-

Google Gemini
Google Gems
NotebookLM
CoPilot
Apple Intelligence

Appendix 2: Development of Gen AI Literacy for Staff and Students

Gen AI Literacy for Staff

Across the three categories of staff use of Gen AI tools, there is an increasing level of Gen AI literacy required.

Gen AI Literacy for Fundamental Staff Use

In addition to understanding which Gen AI tools to select and the appropriate data that could be used, all staff need to have a basic appreciation of the limitations of many Gen AI tools including their potential to produce hallucinations, to produce biased or out-of-date content, and their use of all data inputs to further train the Large Language Model that powers them.

Gen AI Literacy skills for Fundamental Staff Use include the ability to:

verify and review all Gen AI outputs particularly to identify inaccurate or biased content.
discern endorsed Gen AI tools, and other tools and the appropriate data input for the tool.
create basic prompts to achieve desired outcomes.
understand the ways students are using AI.

Fundamental Gen AI Literacy questions include:

Is this output free from discriminatory or harmful content?

Is this output true, accurate, and up-to-date?

Is this an appropriate use of AI with my peers or who I am communicating with?

Is the data I am using appropriate for the tool I'm using? Is it creating a security or privacy risk?

How can I mitigate the risk this output contains discriminatory, biased or harmful content?

Do I know when and how my students are using AI?

How can I tell if it’s helping them stay engaged or if they’re becoming over-reliant on it?

Gen AI Literacy for Intermediate Staff Use

Intermediate-level (medium-risk) use is characterised by AI-generated content delivered to students through a teacher. Due to the potential far reaching impact, staff require deeper Gen AI literacy skills when working at this level. When using AI content with students, staff must apply additional scrutiny, carefully verifying its quality, accuracy, and appropriateness.

Additional Gen AI Literacy skills at this level include:

The ability to scrutinise Gen AI outputs to ensure their alignment to educational purposes (e.g. curriculum resources or instructional content),
The ability to iteratively create prompts to maximise desired outcomes.
The skill of openly and clearly communicating the use of Gen AI when generating and distributing material.

Intermediate Gen AI Literacy questions include-

Is this output relevant and suitable for our curriculum and the age and stage level of students?

Is this output coherent, and consistent?

How pedagogically sound and curriculum-aligned is the content, activities and/or assessment items?

Can human modifications improve this output and how?

What is the appropriate acknowledgment of AI's contribution to the material produced?

Do I feel confident enough with AI to actually teach my students how to use it responsibly?

Do I feel like I have the right tools to step in and redirect them when they’re using it wrong?

Gen AI Literacy for Advanced Staff Use

Higher risk activities should only be undertaken with CEDoW Endorsed tools. The higher risk activities require additional caution and extensive scrutiny, however they also offer additional benefits.

Additional Gen AI Literacy skills when using Gen AI tools to inform evaluations of students' work include-

Recognising that Gen AI evaluation outputs are informative suggestions and should not be taken as final, conclusive judgments.
The ability to use Gen AI with well designed rubrics and worksamples and align Gen AI evaluations with professional judgement.
Crafting prompts carefully to avoid cultural cues that may introduce bias into the results.

The ability to recognise when Gen AI outputs could distort teacher evaluation.

Additional Gen AI Literacy skills, when purposefully using Gen AI tools directly with students within school learning activities, include the ability to:

thoroughly test the operation of a Gen AI tool and evaluate Gen AI outputs for bias, hallucination, out-of-date information, cultural appropriation etc…
frame the use of Gen AI within an ethical framework.
recognise when Gen AI outputs could harm student wellbeing.
ensure equitable access to Gen AI tools for student use.
avoid activities which de-skill students in critical thinking and logical reasoning.
manage issues such as offboarding and student overreliance upon Gen AI tools.
utilise Gen AI to assist learning, not as a replacement for learning.
discern if Gen AI use could disadvantage specific groups within classes/cohorts.

Advanced Gen AI Literacy questions include-

Is the evaluation accurate? Does it provide sufficient insight into my students? Can I apply my professional judgment to ensure the best outcomes for their learning?
Is the time saved by using Gen AI sufficient and what benefits can I bring through effective use of my time?
What effective pedagogical practices could be implemented in a higher quality way with Gen AI Assistants, tools and how could these help me know that my targets for learning are being met?
Am I aware of my students' Gen AI use, can I effectively judge impact to their wellbeing and engagement in learning and potential overuse?
Do I feel empowered and equipped to guide, educate and direct students AI use?

Appendix 3: Gen AI Literacy guidance: basic resources

Add Link to CEDoW AI Site coming soon.

What Is Generative AI - A CEDoW AI Insight?

Accuracy and hallucinations

When asking for facts about a person or topic Generative AI makes up answers (called hallucinating) anywhere up to 5% of the time.

A simple way to demonstrate this is as follows-

Errors - Make LLM incorrectly solve a maths problem by telling the answer first.

Hallucination - Ask LLM for quotes for a novel but suggest a character that does not exist.

See Additional Resources for examples of how to evaluate AI outputs.

Biased Content

Gen AI outputs are built on biased data and may not ever be fully corrected. The bias can be difficult to detect due to subtleties in the underlying assumptions in the model or the expression of them in generated content.

A simple way to demonstrate this is as follows-