The reason computers were developed in the first place was because people thought if you could make enough calculations, you could predict the weather accurately. But like they said in Jurassic Park, if a butterfly farts in Seattle, it rains in Central Park instead of being sunny. There just couldn’t be enough data collected to make reliable predictions. The same is true for the design of tailings facilities, but in a different way. For example, a parameter called permeability can vary by at least 15 orders of magnitude.
Natural materials aren’t predictable. Every time I try to outguess Mother Nature, she makes a fool of me.
Still, I’ve read that computers are more accurate than doctors at diagnosing illnesses (this still doesn’t mean to search WebMD to see what’s ailing you). Computers are relying on data and statistics while doctors may have biases or other matters that misdirect them. I suppose I could replace the word “doctors” with “engineers” in that previous sentence.
All that being said, I don’t think that AI will take over from geotechnical engineers designing, surveilling and watching over the construction of tailings facilities, but we should sure be on the look out for where it can assist us.
I can sure think of a few things that could be mechanized, but you could never replace the judgment of an experienced engineer for some of those things.