AI for Engineers

Yes, use ChatGPT to support a calculation check, not to replace independent verification. Ask it to work through the supplied steps and identify assumptions or possible unit inconsistencies. Rerun important calculations using your established methods and check any cited standard against the original text before relying on the response.

Good prompts specify the engineering task and inputs, request a clear output format, and ask for assumptions to be stated. This hub includes tested prompts for calculation checks, technical reports and standards lookup, with dated, rubric-graded runs. Adapt a prompt to your project and independently verify its output rather than treating a test run as a guarantee.

Choose an AI tool by testing it on your engineering task and reviewing its data handling, rather than relying on a universal ranking. Check enterprise or no-training settings alongside storage and access arrangements. Compare its output against work you have already verified, using only material you are authorized to share.

Do not assume it is safe; permission to share and the tool's data handling both need review. Check your client's terms and your employer's policy before entering confidential material. Review retention and access controls as well as training settings, and use a nonconfidential example if authorization or handling arrangements are unclear.

Under the NCEES Model Law (revised August 2025), the licensee who seals the work remains responsible for it, whatever tool produced the draft. The model law defines Responsible Charge as exercising full professional knowledge of and control over the work. Verify the rules applicable to your work with your state PE board.