The Gartner Hype Curve applies here. It's actually a great technology, but not as great as execs are claiming. So there will be a crash followed by s a slow climb to productivity and eventually standard use for significant gains in efficiencies.
I'm in an industry that really wants to use AI, but we're highly regulated and it's a little like pounding a square peg into a round hole from a development assurance perspective. Our processes rely on human authors for requirements and design as well as independence layers built in. If a person is using AI to author requirements for any product, how do we substantiate where Ai stopped and the human started? What kind of defects are we now looking for if this paradigm has changed?
I think there's a lot more of that stuff built into regulations for all sorts of products people aren't taking into consideration properly right now. From medical, to automotive, to aerospace and many other industries, human authorship and independent human verification is built into the regulatory basis for hundreds of thousands of products. Are you building heart valves or shower valves? Are you building an app for a phone that helps shoppers buy groceries or an app for a flight management computer that helps pilots account for winds in calculating fuel burn?
And even things like accounting, human input is assumed in a lot of aspects of GAAP. There are independence assumptions in GAAP too. If AI is checking its own work, you don't have independence.
It all sounds great, and might be working towards capable, but you've still got millions of pages of regulations for the production of hundreds of thousands of products to take into consideration.
= new reply since forum marked as read