12 principles for the ethical, safe, and responsible use of generative artificial intelligence.
Do not share confidential information in public AI tools
Before using tools like ChatGPT, Gemini, or other external platforms, ask yourself: Is this content internal, strategic, or confidential? If the answer is yes, do not enter it. Remember that we have approved corporate tools that protect our information.
Always verify the information generated by AI
AI can make mistakes or present incorrect data with apparent confidence. Before making decisions or sharing results, cross-check the information with reliable sources and apply your professional judgment.
Consult before using AI in sensitive or critical decisions
If you are going to rely on AI for decisions that impact people—such as hiring, performance evaluation, health, or safety—consult in advance with your leader or with the DNA Laboratory. These decisions require human oversight.
Participate in AI training sessions
The ethical, safe, and effective use of these tools requires ongoing training. We invite you to participate in the sessions offered by the DNA Laboratory, La Academia, or the People area.
Label AI-generated content
When you use AI to create texts, images, presentations, or analyses, disclose this transparently. For example: "Content developed with the support of artificial intelligence."
Do not use AI to generate false, offensive, or manipulative content
It is prohibited to use AI tools to create misinformation, impersonate identities, or generate deepfakes. This type of use may have disciplinary consequences.
Identify and question biases in AI responses
If an AI response appears to unfairly favor a person or exclude a group, do not use it as is. Adjust it, question the result, and if you deem it necessary, validate it with your leader.
Do not enter personal data from third parties
Avoid entering into AI tools personal information about clients, suppliers, or employees, such as names, identity documents, emails, or sensitive data. Improper handling of this data can create legal implications for the company regarding data protection and regulatory compliance.
Remember that you are responsible for the final result
Using AI does not transfer responsibility. If a text, analysis, or decision contains errors due to blind trust in the tool, the responsibility lies with the person who used and validated it. Always review, adjust, and rely on your own judgment before sharing or acting on any AI-generated output.
Verify that the tool does not use your data for training
Before using any AI tool, check whether it can be configured so that the information entered is not used to train its models. If in doubt, opt for the approved corporate tools.
When in doubt, ask
If you are not sure whether a tool is safe or permitted at Argos, consult with the DNA Laboratory or with Summa. It is better to ask before acting.
If you create an AI agent or assistant, do it responsibly
More and more collaborators can create their own AI agents or assistants to automate tasks. If you are going to build one, keep three key things in mind: first, be mindful of what data you are training or feeding it with—if you use confidential, strategic, or personal data, that agent becomes a sensitive repository that must be protected as such. Second, define who can use it—a poorly configured agent can be accessed by unauthorized people inside or outside the organization, which represents a real risk of information leakage. Third, before deploying it, validate with your leader and your team who should have access, especially if the agent handles confidential information. If at any point you need guidance on how to do this safely, the DNA Laboratory and Summa are available to support you.
