What "Responsible Charge" Means for AI-Assisted Design
Responsible charge requires active engineering involvement during the design process— not just review of the finished product. For AI-assisted work, that means human oversight while the AI is generating outputs. Not a rubber-stamp approval after the fact.
This distinction matters. According to professional liability analysis1, "reviewing or correcting technical submissions after they have been prepared and completed by another designer does not constitute responsible control." Replace "another designer" with "an AI tool" and the principle holds exactly.
The NSPE Board of Ethical Review made this explicit6. In a case evaluating an engineer's use of AI, the board found that "Engineer A's misuse of the tool, by failing to maintain Responsible Charge over the AI tool and its output before sealing the document, was unethical." Not ambiguous.
Think of AI like an intern. A talented one— maybe Ivy League. But if that intern makes an error in a design calculation, the supervising PE is accountable. Not the intern. Structure Magazine's analysis7 confirms this parallel: "If AI contributes to a design error, it is comparable to mistakes made by interns or junior engineers." The tool doesn't absorb your responsibility. You do.
As the American Society of Civil Engineers (ASCE) states in Policy Statement 573, "AI cannot serve as a replacement for the professional judgement of a licensed Professional Engineer"3. The professional judgment is yours. The AI is the tool.
Responsible Charge: What Counts vs. What Doesn't
| Meets Responsible Charge | Does NOT Meet Responsible Charge |
|---|---|
| Directing AI parameters and constraints for each design task | Accepting AI recommendations without validation |
| Reviewing intermediate calculations at each design step | Post-hoc review of final AI output only |
| Making engineering decisions based on AI output evaluation | Rubber-stamping AI-generated documents |
| Verifying AI outputs against applicable codes and standards | Assuming AI "knows" the relevant building codes |
| Documenting your engineering rationale at each stage | No documentation of AI use or validation steps |
What ASCE, NSPE, NCEES, and NCSEA Actually Say
Every major engineering standards body— ASCE, NSPE, NCEES, and NCSEA— has taken the same position: AI is a permissible tool, but it cannot replace professional engineer judgment. The engineer's legal and ethical obligations remain exactly the same.
There is no dissent here. Every organization says the same thing— here's exactly what each one requires:
Standards Body Positions on AI in Engineering
| Organization | Key Position | Date | Emphasis |
|---|---|---|---|
| ASCE (Policy 573) | "AI cannot serve as a replacement for the professional judgement of a licensed Professional Engineer"3 | July 2024 | Public safety priority; rapid AI advancement outstrips current laws |
| NSPE (Position 03-1774) | Same licensure standards apply to AI oversight; requires verification, validation, continuous monitoring4 | 2024 | Rigorous testing, transparency, accountability |
| NCEES | "Engineers or surveyors using AI must retain ultimate responsibility for all decisions"5 | December 2025 | Clear documentation, continuing education, verifiable outputs |
| NCSEA | "AI use does not change a licensed engineer's legal responsibilities"2 | 2024 | Building code compliance, anticipates lawsuits from AI failures |
ASCE's Policy Statement 573, adopted July 18, 2024, sets the clearest marker. It acknowledges that "rapid advancement of AI technology is outstripping current laws and regulations"3 while affirming that the PE's core obligations don't shift. NSPE goes further— requiring that individuals who "design, develop, or oversee AI systems that have a direct impact on public safety should be held to the same professional licensure standards as traditional engineers"4.
The NCEES position, published in December 2025, adds the documentation requirement5. AI outputs must be verifiable with clear documentation of methodologies, data sources, and assumptions. Licensees must obtain continuing education on AI capabilities and limitations.
Use AI if it helps you work. But don't pretend it changes what your stamp means. An AI governance strategy that accounts for these standards body positions should be foundational to any engineering firm's AI adoption plan.
The Insurance Problem— E&O Coverage Gaps and AI Exclusions
As of January 1, 2026, major insurance carriers have added AI exclusions to professional liability policies for design professionals. If your firm uses AI in any design work, your errors and omissions (E&O) coverage may no longer protect you.
This isn't hypothetical. It's already happening.
Verisk— the insurance industry standard-setter— released standardized AI exclusion forms CG 40 47 and CG 40 48, effective January 1, 20268. Major carriers including Berkley, AIG, and Great American are adopting these exclusions for professional liability policies8. The exclusions specifically target losses arising from tools including ChatGPT, Midjourney, and DALL-E8.
Research cited in insurance industry analysis suggests general-purpose AI tools produce inaccurate outputs— known as hallucinations— at rates between 58% and 88%8. For design professionals, that level of unreliability translates to potential structural violations, specifications for nonexistent materials, and calculations that fail engineering standards.
Traditional E&O policies weren't built for this. Armilla AI— a company authorized to write insurance policies through Lloyd's of London— identifies the core gap: professional services firms need "affirmative AI coverage that explicitly includes 'professional services delivered or assisted by AI'"9. Standard policies simply don't address it.
The emerging solution is purpose-built AI insurance. Armilla's products cover "financial damages and legal defense costs related to an AI model's underperformance"9— including model errors, hallucinations, and inaccurate outputs. It's early, but the market is responding.
Insurance Landscape for Engineering Firms Using AI
| Action | Why |
|---|---|
| Audit current E&O policy for AI exclusion language | CG 40 47 and CG 40 48 may already apply to your policy |
| Ask your broker about affirmative AI coverage | Traditional policies won't cover AI-related claims |
| Document all AI tool usage in engineering workflows | Creates a defensible record of responsible charge |
| Disclose AI use to your insurer proactively | Non-disclosure could void coverage entirely |
The hidden costs of AI projects aren't just about implementation budgets. Insurance gaps are the surprise most firms aren't planning for.
State Board Guidance— Where Things Stand
Texas was the first state to issue clear guidance on AI use by professional engineers. Most other states haven't caught up— leaving firms to fill the regulatory gap using national standards body positions.
Texas PELS Policy Advisory Opinion 71, approved November 14, 2024, permits AI use by licensed engineers with a straightforward caveat10:
"AI software should not be relied upon without oversight of engineering and surveying concepts and principles."
That's as clear as state-level guidance gets right now. No specific regulations in the Texas Engineering Practice Act directly address AI— PAO-71 fills the gap through advisory opinion.
NCSEA confirms the broader picture: "There are presently no specific regulations exclusively governing AI use by professional engineers"2. That gap is the problem. Without state-specific rules, firms are making their own judgment calls about AI use— and when something goes wrong, those judgment calls will be evaluated against the national standards body consensus.
Other major states— California, New York— have not issued specific guidance as of April 2026. The practical implication: firms should treat the national standards body consensus plus the Texas model as their baseline. More state boards will follow. The firms that have frameworks in place when those regulations arrive won't be scrambling. Everyone else will.
How to Use AI Without Risking Your PE License
Using AI safely in engineering practice comes down to three things: maintaining responsible charge throughout the design process, documenting everything, and auditing your insurance coverage. Here's the framework.
1. Maintain responsible charge throughout the process. Direct the AI. Validate intermediate outputs. Make engineering decisions at each step— not just at the end. "Reviewing or correcting technical submissions after they have been prepared and completed" doesn't count1.
2. Document AI use comprehensively. Record which AI tools were used, what tasks they performed, how outputs were validated, and what changes you made based on professional judgment. NCEES requires that AI outputs be "verifiable with clear documentation of methodologies, data sources, assumptions"5.
3. Implement organizational QA/QC processes. This isn't optional. Structure Magazine puts it directly: "From an organizational standpoint, you must have a QA/QC process to use this tool"11.
4. Validate against engineering standards. AI-generated designs must "comply with all relevant building codes and meet safety and performance requirements"2. The AI doesn't know your jurisdiction's code requirements. You do.
5. Audit your E&O coverage. Check for AI exclusions. Consider affirmative AI coverage. Firms need coverage that explicitly addresses professional services delivered or assisted by AI9.
6. Distinguish AI-assisted from AI-generated. Use AI for research, documentation, code review, and initial calculations. Maintain human judgment for all design decisions. AI can genuinely accelerate documentation, code review, and preliminary analysis— the value is real. The question is whether your oversight keeps pace with your output. As Structure Magazine warns: "Do NOT rely on ChatGPT or AI tools to provide accurate outputs. You must have the expertise to vet results for accuracy"11.
7. Keep current. NCEES requires continuing education to "ensure successful AI integration while upholding best practices"5. The technology changes fast. Your understanding of it needs to keep pace.
A clear AI decision framework helps determine where AI adds value and where professional judgment remains irreplaceable. And measuring AI success in your workflows ensures that efficiency gains don't come at the cost of professional standards.
FAQ— PE Stamp and AI
Can a professional engineer stamp AI-generated designs?
Yes— if the engineer exercises responsible charge and validates all outputs. But the engineer assumes full personal and professional liability for any errors, the same as if they had created the design themselves13. The stamp means you own it. All of it.
Does using AI reduce a PE's professional liability?
No. Every major standards body— ASCE, NSPE, NCEES, and NCSEA— confirms that AI use does not change a licensed engineer's legal responsibilities2345. The PE stamp carries the same weight regardless of how the design was created. (If anything, AI use invites more scrutiny, not less.)
Will my E&O insurance cover AI-assisted engineering work?
It depends on your policy. As of January 1, 2026, major carriers including Berkley, AIG, and Great American have added AI exclusions to professional liability policies8. Firms should audit current policies and consider affirmative AI coverage products9.
What documentation do engineers need when using AI tools?
Document which AI tools were used, what tasks they performed, how outputs were validated, what changes were made based on professional judgment, and the QA/QC process followed511. This demonstrates responsible charge and creates a defensible record.
Are there state regulations about engineers using AI?
As of April 2026, Texas has issued the clearest guidance— PAO-71, approved November 2024— permitting AI use with engineering oversight10. No specific regulations exclusively govern AI use by PEs in most states2. National standards body positions serve as the primary framework.
The Window Is Now
Engineering firms that build responsible AI practices now will be better positioned when the first major liability cases arrive. And NCSEA is clear: those cases are coming2.
The firms that establish frameworks today— maintaining responsible charge, documenting workflows, auditing insurance coverage— will have the documentation, the processes, and the protection to weather them. The ones that don't will be assembling their defense after the fact.
More state boards will follow Texas. Insurance carriers will refine their positions. But the core principle won't change: AI is a tool, and the engineer who stamps the work is responsible for the result. That clarity is actually freeing— it means you can explore what AI does well without ambiguity about where the line is.
Just because AI makes engineering outputs easier to produce doesn't mean those outputs are good. That judgment— the difference between easy and good— is exactly what the PE stamp represents.
If your engineering firm is navigating AI adoption and needs to understand both the technology opportunities and the liability landscape, an AI strategy for professional services is the place to start.
References
- Cavignac & Associates, "Signing, Stamping and Sealing Others' Designs" (2024) — https://cavignac.com/blog/signing-stamping-and-sealing-others-designs/
- NCSEA, "AI in Structural Engineering - Q&A" (2024) — https://www.ncsea.com/foundation/innovation/artificial-intelligence-in-structural-engineering/ai-townhall-questions-answers/
- ASCE, "Policy Statement 573 - Artificial Intelligence and Engineering Responsibility" (2024) — https://www.asce.org/advocacy/policy-statements/ps573---artificial-intelligence-and-engineering-responsibility
- NSPE, "Artificial Intelligence Position Statement 03-1774" (2024) — https://www.nspe.org/nspe-advocacy/explore-issues/professional-policies-and-position-statements/artificial-intelligence
- NCEES, "Licensure Exchange December 2025, Vol. 29 No. 6" (2025) — https://ncees.org/december-2025-licensure-exchange/
- NSPE, "Board of Ethical Review - Use of AI in Engineering Practice" (2024) — https://www.nspe.org/career-growth/ethics/board-ethical-review-cases/use-artificial-intelligence-engineering-practice
- Structure Magazine, "Transforming Structural Engineering: Embracing the AI Revolution" (2024) — https://www.structuremag.org/article/transforming-structural-engineering-embracing-the-ai-revolution/
- Verisk / Insurance Industry, "Insurance Carriers Add AI Exclusions to Design Professional E&O Policies" (2026) — https://www.financialcontent.com/article/marketersmedia-2026-1-16-insurance-carriers-add-ai-exclusions-to-design-professional-e-and-o-policies
- Armilla AI, "AI Insurance Brief" (2025) — https://www.armilla.ai/ai-insurance-brief
- Texas PELS, "Policy Advisory Opinion 71 - Use of AI Software by Licensees" (2024) — https://pels.texas.gov/nm/2024/pao-71-response.pdf
- Structure Magazine, "ChatGPT in Structural Engineering" (2024) — https://www.structuremag.org/article/chatgpt-in-structural-engineering/
