AI Will Become a Licensed Professional Engineer
Happy Friday,
AI will become a licensed Professional Engineer, with its own stamp, and it will be irresponsible not to use it.
We license engineers because engineering mistakes can hurt people. Licensure establishes a minimum level of competence, judgment, and ethical responsibility before someone is entrusted with decisions affecting public health and safety.
Now imagine an AI that can pass the PE exam.
Actually, forget today's exam. Make the AI exam much harder.
Require near-perfect performance across thousands of engineering problems. Test every obscure provision of every applicable code. Give it incomplete information and see whether it recognizes what's missing. Test competing design alternatives, uncertainty, constructibility, economics, environmental consequences, and engineering ethics.
Then test it again next week.
Compare that with a human engineer who passed the PE exam 20 years ago, perhaps by a single question, and hasn't been formally tested since.
Which should the public trust?
AI will consider more variables simultaneously than any human could, absorb vast bodies of technical knowledge, and learn from failures across thousands of projects. It won't get tired, distracted, rushed, or have a bad day.
And when the data begin showing that AI-assisted engineering produces fewer costly errors and claims, professional liability insurers will notice.
Imagine the premium schedule:
Human-only engineering: $$$$$$
AI-assisted engineering: $$$$
Licensed-AI engineering: $
Market forces and growing public trust in AI will drive the change. Insurers will reward firms that reduce risk. Clients will reward firms that reduce their cost and risk. And these will compel engineering firms to employ licensed AI to remain competitive, i.e., to survive.
Shouldn't a human engineer check the AI's work? Do we ask human engineers to recheck their calculators and spreadsheets by redoing the arithmetic by hand? No, because calculators and spreadsheets have proven correct a statistically significant percentage of the time. The trajectory of AI implies it'll get there too.
An analogy: As autonomous vehicles become dramatically safer than human drivers, saving many lives, society won't insist humans keep driving simply because driving was historically a human activity. On the contrary, driving a two-ton projectile around other people with just two eyes and comparatively slow reflexes will be seen as a selfish, irresponsible act."My child isn't going if your husband insists on driving..." Drivers will be shamed like a smoker in a restaurant. Engineers won't be shamed, but you get the point.
Punchline: Professional engineering licensure was created to protect the public, not engineers. When AI protects the public better than humans do, the question won’t be whether to license the AI. The question will be whether it's responsible to continue licensing humans. Market forces will make it inevitable. Sometimes you just have to ride a horse in the direction it's facing.
Have a great weekend,
Dave
P.S. I feel I should add that this is not as dire as it seems. There are strategies firms and individuals can employ now to prepare and capitalize on the changes coming to our profession. Put on your entrepreneur hat. See the big picture.
Feedback and blowback are always welcome: dave@goodnewsfriday.com
All previous topics are available at @goodnewsfriday.com
Good News Friday: Successful Engineering Consulting, both hardback and Kindle, is on Amazon. 😄
Member discussion