This article discusses artificial intelligence text models — programs that can read your natural language requests and generate texts, tables, lists, or answers to questions. They differ in purpose and complexity:
Chatbots — answer questions and maintain a conversation.
Text generators — can write articles, instructions, and resumes.
Specialized models — trained on a specific field, such as medicine, law, or HSE.
To dive in, we will use a universal text model, such as ChatGPT, because such a model:
Understands ordinary language.
Is capable of holding a conversation, answering questions, and creating instructions.
Is reasonably safe and accessible via free apps or a browser.
The goal is to learn step-by-step how to use a text model as a personal assistant, and then gradually train it to perform HSE tasks.
Open a free app on your phone or use a browser. Ask any questions to understand how the AI responds.
Examples:
"Find a cozy seaside hotel in Pattaya and explain why it's a good choice"
"How to wish a colleague a happy birthday so it sounds sincere and original"
"Write a short birthday greeting in English"
Create a separate tab or document for each request. This way, you can return to successful solutions and gradually build your personal prompt library.
Tip: At this stage, the main thing is to get used to the format and see that the AI understands ordinary language and does not require technical skills.
The AI does exactly what you ask. For the answers to be useful:
Provide context: "You are an HSE specialist at a large enterprise."
Clearly state the task: "Create an instruction," "Explain in simple words."
Define the response format: list, table, short paragraph.
Examples:
"You are an HSE expert. Explain in simple words why an HSE management system is needed"
"Create a table: hazard — control measure — responsible person"
"Prepare a short birthday greeting for a colleague"
Tip: When you find a successful solution, save it in a separate tab. This builds your personal prompt library, which can be reused.
After mastering simple requests, you can move on to HSE work:
Find current requirements for employee training.
Compare PPE standards for office and production employees.
Create an incident report template.
Explain the difference between a safety briefing and on-the-job training in simple words.
Use tables or lists so the results are immediately convenient for work. It is better to create a separate tab or document for each type of task.
Tip: In the beginning, use simple data — regulatory documents, internal instructions, short reports. This will allow you to practice with the AI without risk.
When simple tasks are mastered, you can gradually increase the complexity:
Analyze documents and highlight key employer responsibilities.
Summarize long regulatory documents to one page without losing meaning.
Create incident probability tables by risk categories.
Compare ISO 45001 requirements with national standards.
Tip: Keep separate tabs for each type of complex task. This way, you can quickly find the best solutions and form ready-made templates for future work.
Save every successful request. Divide the library into categories, for example:
Analysis of regulatory documentation
Document preparation
Reporting and presentations
Risk analysis
Communications and training
Put only truly working solutions and requests into the library. Over time, this turns into a set of ready-made tools that can be used over and over again.
Tip: Do not try to save everything. Save only prompts that give a high-quality result on the first or second try.
When the previous steps are mastered, you can assign more complex tasks:
Selecting measures to reduce risks.
Checking instructions for compliance with requirements.
Creating dashboards for safety indicators.
Preparing reports and presentations for management.
Thus, your AI gradually turns into a full-fledged assistant that helps solve professional tasks and frees up time for the main thing — risk analysis and working with people.
Start with simple questions, gradually complicate tasks, save the best solutions in a prompt library, and reuse them.
In a few weeks, you will have a personal assistant who saves time, systematizes information, and helps make decisions.