AI Literacy to Learn Without Losing Your Critical Thinking Skills
AI literacy is the knowledge and skills we need to understand artificial intelligence, use its tools and critically assess their outputs and implications. It helps us decide when AI might be useful, how to check its work and when we’re better off without it.
Through my experiments and training sessions, I’ve encountered very different reactions. Some people are worried, others are angry, and many focus mainly on the possibilities. I tend to fall into the enthusiastic camp. Still, I believe the quality of our questions and the care we take to verify answers deserve just as much attention as the tools’ capabilities.
AI Literacy Goes Beyond Writing Prompts
Knowing how to write a good prompt is useful. But we also need to assess the response, understand the tool’s limitations and decide whether it’s appropriate for the task.
Université Laval’s Académie de la transformation numérique uses animal profiles as a learning tool to help people understand their current level of AI literacy. Which one are you? 🦊 If you’re not sure, the Académie offers a self-assessment tool to help you assess and improve your AI literacy. These profiles bring together three dimensions:
✅ understanding the concepts;
✅ knowing how to use the tools; and
✅ thinking critically about their implications and limitations.
This approach gives people a starting point for improvement, whatever their current level. In practice, these dimensions overlap. If I ask an assistant to summarize a document, I need to understand what I’m sharing with it, compare its summary with the original and decide whether the result is faithful enough for its intended purpose. Fluent wording doesn’t guarantee accuracy.
Understanding Different Reactions to Open Up the Conversation
People who criticize AI often raise important questions about intellectual property, how models are developed and changes to the way we work. Dismissing all these objections as simple resistance to change would do little to move the conversation forward.
Fear also deserves to be heard. During a ChatGPT training session, a philosophy professor sharply criticized me for supporting a technology he believed was harmful to the next generation. I didn’t share his conclusion. Still, his reaction expressed a concern we need to take seriously: what happens to learning when a tool can produce an answer for us?
At the other extreme, enthusiasm can make us less alert to mistakes. I know I sometimes tend to see the possibilities first. That outlook encourages me to experiment, but it needs to be balanced with careful checking and a willingness to acknowledge what isn’t working.
Then there are people who have used an assistant to rephrase a message or come up with an idea without exploring much further. What they need most is accessible guidance on what they can ask it to do and what they should check.
Starting with the Need Before Choosing a Tool
To develop responsible AI habits, I suggest starting with a task you understand well enough to assess the result. Preparing an outline, comparing two ways of wording something or organizing notes can already prompt useful questions.
Before experimenting, identify what you want to improve. Are you trying to clarify an idea, save time or consider another perspective? Having a clear goal helps you judge the result on more than how impressive it looks.
You also need to consider the information you plan to share. When you’re learning, a fictional example or a public document will often be enough. In a workplace or educational setting, approved tools and applicable guidelines should be part of the process from the start.
Info IA Québec’s guides and sample prompts can provide starting points for these experiments. Prompts still need to be adapted to the context, and the results need to be checked.
Checking AI Responses Against Independent Sources
An AI response can contain mistakes even when it seems clear and well structured. Asking the tool whether it’s sure of its answer is no substitute for checking the relevant documents or sources.
For a summary, compare the key facts with the original text. For a calculation, check the inputs and redo the calculations. For information attributed to an organization, I would consult that organization’s publication rather than rely solely on the link or quotation provided by the AI. A few questions can help guide this review:
The level of scrutiny should reflect the intended use. A draft headline that will be revised and a document that will inform a decision don’t require the same level of review.
Using AI to Support Learning
In education, cheating features prominently in discussions about AI. I tried to introduce AI training through my college’s complementary course offerings, but wasn’t able to move the proposal forward. That experience led me to reflect on how much attention we give to teaching people to use AI alongside setting rules for its use.
Allowing a tool doesn’t mean people know how to use it. Students and professionals alike need opportunities to experiment, compare results and explain their approach.
For example, someone could first write out their own reasoning, then ask an assistant to identify its weaknesses. The next step would be to assess the assistant’s objections and decide which ones are worth considering. Another exercise could involve comparing an AI-generated response with reliable sources and explaining the corrections needed.
In these examples, the person remains actively involved in the thinking. They need to understand what they keep and be able to justify their choices. Getting a piece of writing done faster doesn’t, on its own, demonstrate that learning has taken place.
Learning with Discernment and Being Open About Limitations
I remain optimistic about AI’s possibilities. I’ve seen it help people structure ideas and explore different ways of working. I also want those who speak positively about AI to share their mistakes and explain when they prefer to use another approach.
Exchanging ideas with other users can help us reflect on our habits. An AI community of practice, for example, provides a setting where people can compare approaches and discuss how they check their work.
Developing AI literacy means learning to exercise judgment in these situations. I want to pass on that ability as much as my interest in the tools: being able to experiment, recognize our limitations and remain accountable for how we use the results.
Would you like clearer guidance on using AI and practical methods that fit your needs? Explore my coaching and training options on my website, or contact me at info@jimmygilbert.com to discuss your needs and goals.
Jimmy Gilbert, Technology Consultant and Trainer. A computer science instructor at Cégep de Sainte-Foy, Jimmy brings together expertise in teaching, software development and artificial intelligence. He helps prepare the next generation of technology professionals by combining technical rigour with a practical understanding of industry needs, while actively incorporating AI tools into his teaching. He also works as a technology coach, helping individuals and professionals make effective use of technology, automation and AI.
