When someone asks me what Ethiceye does, the shortest answer is this: We don’t teach how to use ChatGPT, that’s what anyone does. What we do is build with organizations and teachers the ability to decide when to use it, for what, with what limits and what they will never delegate to a machine.
That has a name: criteria. And it is not the same as knowing tools.
The difference seems subtle. it is not.
Knowing how to use ChatGPT does not protect you when CHATGPT changes. Have criteria yes. |
What gives learning tools
Learning how to use AI tools gives you short-term efficiency. You can generate faster text, summarize documents, create talks, automate repetitive tasks. That has value.
But the tools change. ChatGPT today is not the same as there were twelve months ago. The one that will be in twelve months does not look like this. The functionalities, the limits, the models, the interfaces, everything changes continuously.
What you learned about a specific tool expires. And if what you have is just tools knowledge, you have to relearn every time something changes. Which, in AI, is practically constant.
What gives build criteria
The criteria is different. It doesn’t depend on what tool you use or how it works technically. It depends on questions that are still valid regardless of the tool:
For what exactly do I use this AI in this context? What data does it process and what consequences does that have? What biases can you have in this specific case? What do I check before using your output? What decisions will I never delegate to a machine?
Those questions work with ChatGPT, with Gemini, with Claude, with CoPilot or with the tool that exists in three years. The criteria does not expire because it is not anchored to a specific technology, it is anchored to its own judgment.
Why this matters in organizations
I have entered organizations that had done extensive training on AI tools (prompting workshops, Gemini courses, demos of the latest developments) and yet could not answer basic questions: Who has decided that the team can use these tools with customer data? What happens If someone generates a report with AI and contains an error? Who answers?
Tool knowledge does not answer those questions, the criteria does.
An organization with criteria can incorporate a new tool in days because it already knows how to evaluate it. You can answer to a client asking how you use AI because it has a documented position. You can manage an incident because it has a process.
An organization without criteria learns the new tool, but still cannot answer the questions that matter.
Why this matters in education
The same applies in educational centers, with an additional implication: what teachers do with AI is what students learn to do.
A cloister who learns tools transmits to students that AI is an efficient resource. A cloister that has criteria conveys something different: that technology is used with conscience, that there are questions to ask before delegating and that one’s own thinking is not optional.
The difference is not in whether the teacher knows how to use ChatGPT, it is in whether the teacher can explain to a student why some things are not going to be delegated to a machine. That requires criteria, not just technical skill.
What teachers do with AI is what students learn to do. The criteria is taught by example, not with the manual. |
What does it mean to build criteria in practice?
Constructing criteria is not a theoretical training on AI ethics. It is a concrete process that has real deliverables:
In an organization: A policy of using AI co-constructed with the team, which answers who decides, what can be done and who responds if something fails. Something that the team understands and can use tomorrow.
In an educational center: an institutional position on the use of AI, built with the cloister, which gives coherence to the decisions of each teacher and that the center can defend against students, families or an inspection.
In a person: a set of own questions (not Ethiceye, yours) that applies before using AI in your work. who knows what he delegates and what not. that can justify that decision.
That’s what’s left when Ethiceye ends. Not the Prompts manual, the criteria.
The question that sums it up
When someone evaluates if an AI training has helped you, you often wonder if you now know how to use the tools better. It is a reasonable question.
But there is another more demanding question: do you now know what decisions you are not going to delegate to a machine? Can you justify how you use AI before who cares if you trust in you?
If the answer to that question is yes, there has been criteria. If the answer is just “I know how to use better ChatGPT”, there has been a knowledge of tools.
Both have value, but only one hard.
