Mike Howell, PE
EASA Technical Support Specialist
In the rapidly evolving field of electromechanical technology, staying up to date with the latest tools and methodologies is crucial. One such tool that has gained significant traction is the Large Language Model (LLM). These models, powered by artificial intelligence (AI), can process and generate human-like text, making them invaluable for a variety of applications, including technical support, documentation, and troubleshooting. Just remember, these tools can be prone to error and can produce unrealistic or erroneous results as demonstrated in Figures 1 and 2 where ridiculous prompts still produced outcomes.
Understanding Large Language Models (LLMs)
Large Language Models, such as GPT-4, are advanced AI systems trained on vast amounts of text data. They can understand context, generate coherent responses, and even provide detailed explanations on complex topics. For electromechanical technicians, LLMs can be a powerful resource for finding solutions to technical problems, generating documentation, and enhancing communication with colleagues and clients.
Good Practices for Prompt Engineering
A prompt is the text used to ask a question or pose an issue to an LLM. Prompt engineering is the process of designing and refining prompts to elicit the best possible responses from an LLM. Here are some good practices to follow when crafting prompts for technical questions:
- Be Specific and Clear: The more specific and clear your prompt, the better the response. Avoid vague or ambiguous language.
- Provide Context: Include relevant background information to help the LLM understand the context of your question.
- Use Structured Formats: When possible, use structured formats like bullet points or numbered lists to organize information.
- Iterate and Refine: Don’t hesitate to refine your prompts based on the responses you receive. Iteration can lead to better results.
- Test Different Approaches: Experiment with different phrasings and structures to see what works best for your needs.
Examples of Poor and Good Prompts
To illustrate the importance of prompt engineering, let’s look at a few examples:
Example 1: Troubleshooting a Motor
- Poor Prompt: “Why is my motor not working?”
- Good Prompt: “I have a three-phase induction motor that is not starting. It is normally started direct-on-line. It has been working fine for two years, but today it won’t start. It just hums, pulls high current and then trips. I’ve checked the power supply, and it’s fine. What could be the possible reasons for this issue?”
In the good prompt, the technician provides specific details about the motor and the steps already taken, which helps the LLM generate a more accurate and useful response.
Example 2: Generating Documentation
- Poor Prompt: “Write a manual for a machine.”
- Good Prompt: “Create a user manual for the XYZ-2000 CNC machine. The manual should include sections on safety precautions, installation instructions, operating procedures, and maintenance guidelines. Please provide detailed steps and diagrams where necessary.”
The good prompt specifies the type of machine, the sections to be included, and the level of detail required, resulting in a more comprehensive and relevant manual.
Example 3: Understanding a Technical Concept
- Poor Prompt: “Explain how a transformer works.”
- Good Prompt: “Can you explain the working principle of a step-up transformer used in power distribution? Please include details on the primary and secondary windings, the role of the core, and how voltage transformation occurs.”
The good prompt narrows the focus to a specific type of transformer and requests detailed information on key components and processes.
Risks of Relying on AI Responses
While LLMs can be incredibly useful, it’s important to be aware of the risks associated with relying on AI-generated responses:
- Accuracy: AI models can sometimes provide incorrect or outdated information. Always verify critical details from reliable sources.
- Context Understanding: LLMs may not fully grasp the nuances of a specific technical issue, leading to incomplete or irrelevant responses.
- Bias: AI models are trained on large datasets that may contain biases. Be cautious of any biased or inappropriate suggestions.
- Security: Sharing sensitive or proprietary information with an AI model can pose security risks. Avoid disclosing confidential details in prompts.
- Dependence: Over-reliance on AI can reduce the development of problem-solving skills. Use AI as a tool, not a crutch.
Conclusion
LLMs have the power to change the way electromechanical technicians approach problem-solving and documentation. By following good practices for prompt engineering, technicians can harness the full potential of these AI tools to enhance their work efficiency and accuracy. However, it’s crucial to remain aware of the risks and use AI responsibly. Remember to be specific, provide context, use structured formats, refine and adjust your prompts based on responses given, and test different approaches to achieve the best results.
