Navigating Generative AI in Education: A Traffic Light Approach
A Flexible Framework for Integrating AI Tools in your Learning Environment
In a previous post, I wrote about how a scenario-driven approach to developing a generative AI policy or guide in a school or workplace could be a powerful way to consider the range of perspectives when aiming for thoughtful AI use. After running the workshop with a group of educators last week, it seemed as if a flexible approach was required, and an overall policy would not meet the challenges or contribute to experimentation and growth. I suggested that a multi-level traffic light approach could be a good start when aiming to support thoughtful generative AI integration.
An example of what this might look like is below; I also created some icons in Canva, which I will link for download at the end with some other resources.
Level 0: No use of generative AI
This level applies to situations where the use of generative AI is prohibited or not feasible for legal, ethical, technical, or pedagogical reasons. For example, if users' data privacy cannot be ensured or if the generative AI tool is not relevant or appropriate for the learning objectives or the curriculum. At this level, staff and students should refrain from using generative AI and rely on other sources of information and knowledge.
This level could apply when students are taking a standardised test, such as an external NCEA exam. Using generative AI tools to assist on the test is not allowed and could result in serious consequences for the students and the school. Another example could be when marking student work - uploading work to LLMs for training purposes without the student's consent is unethical. Perhaps, in the future, we will not have this level, but for now, and with the way we ‘measure learning’, it might still be required.
Level 1: Limited use of generative AI
This level applies to situations where the use of generative AI is allowed but with restrictions and supervision, for example, if the generative AI tool is experimental, potentially unreliable, or if the users are not familiar or comfortable with the technology. At this level, staff and students should use generative AI only for specific purposes and tasks and only with the guidance and approval of a teacher or educator with expertise in the area. The users should also pay particular attention when verifying the accuracy and validity of the generative AI outputs and accurately cite the sources and methods used (as in all Gen AI use - this is clear on the poster).
An example of a situation where this level applies is when students are learning a new language, such as Spanish or Mandarin. Using generative AI tools to practice their speaking and listening skills, such as Duolingo or Rosetta Stone, is allowed, but only under the supervision of a language teacher or tutor. The teacher or tutor should monitor the student's progress, provide feedback and corrections, and ensure that the generative AI tools are not replacing the human interaction and cultural context of the language learning process. Students should also check the quality and accuracy of the generative AI outputs and acknowledge the sources and tools they used in their learning portfolio.
Level 2: Moderate use of generative AI
This level applies to situations where the use of generative AI is encouraged but with caution and awareness. For example, if the generative AI tool has been tested in the learning context and is considered reliable and beneficial, but it is also established that implementation is complex and potentially unpredictable. At this level, staff and students should use generative AI for various purposes and tasks, but with the understanding and acknowledgement of the limitations and implications of the technology. The users should also evaluate the quality and relevance of the generative AI outputs and provide feedback and suggestions for improvement. Again, at this level, users are encouraged to describe how they have implemented generative AI and even include transcripts or prompts to demonstrate how they have interacted with the tool.
An example of where this level applies is when students write an essay, such as a persuasive or argumentative essay. The use of generative AI tools to generate ideas, outlines, or drafts, such as Essaybot or AI Writer, is encouraged, but with the awareness that the generative AI tools are not perfect and may produce errors, biases, or plagiarism. Students should also critically analyse the generative AI outputs and compare them with other sources and perspectives, such as peer reviews or academic journals. Students should also give feedback and ratings to the generative AI tools and suggest ways to improve them in the future. I have encouraged my students to take this approach with their research idea development here.
Level 3: Extensive use of generative AI
This level applies to situations where the use of generative AI is supported with creativity and responsibility. For example, if the generative AI tool is innovative, trustworthy and reliable for the selected task. At this level, staff and students could use generative AI for multiple purposes, but with the intention and commitment of enhancing the learning outcomes and experiences. The users should also explore the possibilities and opportunities of the generative AI outputs and create original and meaningful work based on them.
For example, students create a multimedia project, such as a podcast or a video. The use of generative AI tools to produce content, such as Descript or Lumen5, is supported, but with the creativity and responsibility of the students to make their work unique and authentic. Students should also experiment with different features and options of the generative AI tools and discover new ways of expressing their ideas and messages. Students should also acknowledge the contribution and influence of the generative AI tools and share their work with others for feedback and recognition. I have encouraged students I am working with to use generative AI to support them in creating summaries of their research - this would be considered to be at 'level 3'.
Level 4: Free use of generative AI
This level applies to situations where the use of generative AI is unrestricted but with respect and ethics. At this level, staff and students should use generative AI for any purpose and task, but with the recognition and appreciation of human values and rights. Again, as will all levels, acknowledgement of generative AI use is best practice.
An example of a situation where this level applies is when students participate in a co-creation project, such as a game or a story. The use of generative AI tools to co-create content, such as AI Dungeon or StoryAI, is unrestricted, but with the respect and ethics of the users to treat the generative AI tools as partners and not as tools. Students should also engage and interact with the generative AI outputs and learn from their responses and feedback. Students should also respect the autonomy and agency of the generative AI tools and not abuse or manipulate them for malicious purposes.
How am I using this approach?
I have just finished writing a new course on research methods, and I have included these guidelines at the beginning and used the icons in any student task so it is clear how Gen AI can be used. There are some research tasks that can be enhanced with Gen AI; however, others aren’t that applicable. I will test the traffic light approach out over this new course and see how it could be improved or if it is even needed at all.
Perhaps an alternative approach to applying this (rather than having the teacher or school dictate how to use GenAI) would be to have students select the level that they thought would be appropriate for the task and then include the icon on the work - EG. L3 - AI with creativity. This would support building AI or digital literacy as well as agency.
Taking it further…
If you are interested in developing these ideas further, you could also use an acronym - I asked Bing Co-pilot to create one (this is something that generative AI is really good at), and this is what it created - I’m sure you could copy this and modify it to suit your school or workplace values.
**S**top: This is the level 0, where the use of generative AI is prohibited or not feasible, and users should not use it at all.
**M**inimize: This is the level 1, where the use of generative AI is allowed, but with strict restrictions and supervision, and users should use it only for specific purposes and tasks, and only with the guidance and approval of a qualified teacher or educator.
**A**nalyse: This is the level 2, where the use of generative AI is encouraged, but with caution and awareness, and users should use it for various purposes and tasks, but with the understanding and acknowledgement of the limitations and implications of the technology.
**R**einvent: This is the level 3, where the use of generative AI is supported, but with creativity and responsibility, and users should use it for multiple purposes and tasks, but with the intention and commitment of enhancing the learning outcomes and experiences.
**T**rust: This is the level 4, where the use of generative AI is unrestricted, but with respect and ethics, and users should use it for any purpose and task, but with the recognition and appreciation of the human values and rights of the technology.
However, I'm not really a fan of the acronym approach and prefer the levels, as the acronym could add another layer of confusion to the guidelines.
Resources and links
Feel free to remix this idea and adapt it to your needs, as I am sure there could be some modifications to improve this approach; you will find a PDF of the poster guidelines here. The icons are in a folder here. If you would like access to the Canvas files to modify, please leave a comment or send me an email, and I’ll share them with you.